<string> is any text line added after the asterisk (*)
The command * (or NOTE) is used for placing comments in a program.
Shark ignores any line that starts with * (or NOTE). Since the
compilation removes all the comment lines, the number of comments has no
effect on the speed of the compiled program.
Notes can also be added with semicolon(;) at the END of any Shark command line. Be careful with the semicolon since placing a semicolon(;) at the BEGINNING of a line can disrupt any following Shark commands!
Example:
In a program: the three comment types are: *,;, and NOTE as shown in the following examples:
* open employee data file:
USE employee INDEX employee ;opening EMPLOYEE data file ordered by name
NOTE these are examples of how to include comments in a program
Assign value to a memory variable, or to a field when a record is referenced
as a vector of fields.
<memvar>=<exp>
<file><num exp>=<exp>
Assigning values to memory variables. The first form of this command
assigns a value to a memory variable; if the variable does not exist, it will
be created. This command is equivalent to the STORE command (see STORE).
Treating a data file as a vector. The second form assigns a value to a
field of a data file. In this form, [ and ] do not indicate an option, [
and ] have to be typed in. For instance, EMPLOYEE[3] is the third field,
ADDR, of the EMPLOYEE data file. This is the only way to assign a value to
a field with the = command. Normally, field values are assigned with the
REPLACE command.
Special care must be taken when using this second form when the data file
is opened with a macro (see USE). Opening a file with a macro allows a
standard program to work with a wide variety of files, provided the compiler
is told the data file structure.
Suppose a data file is opened with the following commands:
USE inven COMPILE
ACCEPT 'Enter name of data file to use: ' to fil_nam
USE &fil_nam
In this case, the compiled program assumes INVEN.DBF is in use, even
though some other file with the same structure is actually opened. Therefore,
if fil_nam='INVEN88', the proper way to reference the 3rd field of the open
data file is inven[3], not inven88[3].
Note that the equals sign, or the REPLACE command, (=) is the only Shark command verb that does not appear
at the start of a command line. All of the following are identical in effect:
1>employee[1]='Steiner' (i.e. replace EMPLOYEE[1] with 'Steiner')
1>REPLACE employee[1] WITH 'Steiner'
1>REPLACE name WITH 'Steiner'
<exp list> the expression or expressions to be displayed
The ? command evaluates the expression or expressions in the list
(remember the commas to separate the expressions in the list!) and displays
the results on a new line. In particular, this command is often used to
display the contents of memory variables and fields.
The displayed items are separated by blanks if SET RAW OFF (see the SET
command; on is the default). When SET RAW ON, they appear side by side.
? with no expression displays a blank line; it is used for spacing on the
screen and printer.
While the ? command moves the cursor to the beginning of the next line
before the items are displayed, the ?? command displays the items at the
current cursor position. The ?? command is used in programs to build lines of
text on the screen or on the printer from several pieces.
Display data and data input request at specific line and column position on
screen or printer, or erase the screen line starting at the specified
coordinates.
EXPRESSION=" "
@ 5,10 say EXPRESSION get VARIABLE PICTURE '999'
READ
line the line number for the display, a numeric expression;
the fractional part (if any) will be discarded
col the column number for the display, numeric expression;
the fractional part (if any) will be discarded
Options:
none erase the screen line starting at the specified
row and column position; when printing, move print head
to specified position
SAY <exp> [USING <format>]
SAY displays the expression, <exp>; if there is a USING clause,
<exp> is displayed using the format specification <format>
GET <var> [PICTURE <format>]
GET asks for an input into the variable <var>; the present
contents of the <var> is displayed; if there is a
PICTURE clause, <var> is displayed using <format>.
The input request is activated by the READ command
Example - To manage printouts:
YESNO='N'
@ 23,18 SAY "Do you want to Print a copy?" get YESNO
READ
IF YESNO ='Y'
SET PRINT ON
LIST
SET PRINT OFF
ENDIF
The @ command is used in Shark programs, first, to display
formatted data at specific locations on the screen or printer, and second, to
prompt the user to type in data (in conjunction with the READ command). Example:
EXPRESSION=" "
@ 5,10 say EXPRESSION get VARIABLE PICTURE '999'
READ
If the SET FORMAT option is TO PRINT (see the SET FORMAT command), then the
(formatted) data is sent to the printer and there can be no GET clause;
otherwise the SET FORMAT TO SCREEN is in effect, and the data is sent to
(and can be obtained from) the screen.
The keywords of the @ command: SAY, USING, GET, PICTURE cannot be in macros.
This command is often used for making input screens and reports in Shark. See the command TEXT that contains many elements of the @ command, and is usually
used in preference to @ for full-screen input and output, and printed output.
If GET <var> is used, the variable, <var> must exist; it is not created by this command. The present value of <var> is shown on the screen; the new
value typed in by the user becomes the contents of <var>; if <var> is a field,
the field in the selected file is changed (i.e. a REPLACE is performed).
If no option is used in screen output, for instance @ 10,0, the line is
cleared on the screen starting at the indicated column. When using a DOS printer, the
print head is moved to the new position.
If there is no <format> clause, the variable is displayed as follows:
fields are displayed with the width specified in the data file structure;
string memory variables are displayed as stored; numeric variables are
displayed as specified by the :PICTURE system variable (see Section 2.6).
The GET clauses are activated by the READ command. There can be no more
than 64 GET clauses for a READ command. The GET clauses with their pictures
are stored in a Get Table, which remains in effect after a READ until another
GET clause is encountered or a CLEAR GETS command is encountered.
NOTES for users of DOS-type printers:
Line numbers go from 0 to 24, column numbers from 0 to 79. For the
printer, the line number counter begins at 0 (line 1); it is reset to 0 by the
EJECT command (see the commands: EJECT, SET EJECT, SET LENGTH TO). If there
are several @ commands for the printer, each @ command must display past the
display of the previous @ command (that is, if the first @ command displays at
line1, col1, and the next at line2, col2, then either line1<line2, or
line1=line2 and col1<col2). If the new print position is less than the current
print position, the printer head is not moved by the @ command.
NOTES for users of Windows-type printers:
Printed output mimics the screen display, and is easily adjusted by "trial-and-error".
ACCEPT ['<string>'] TO <str var>
Option: '<string>' the prompt message
This command is used in Shark programs to request character
information to be placed into a memory variable. The text in <string> will be
displayed as a prompt message. Note that <string> has to be delimited by ' or
by ", and may not be either a macro or a string expression. ACCEPT cannot be
used to input data into a field or element of a matrix.
If the <memvar> does not exist, it will be created.
The optional character string is used as a prompt. A character
expression cannot be used, but a macro is permitted, provided the macro
expression includes quotation marks.
To input numeric or logical data, use the INPUT command. Example In a program:
ACCEPT 'Your name: ' TO name
Your name: David Simco
The following illustrates use of a variable instead of a string as the prompt:
1>prompt='"This is a prompt: "'
1>ACCEPT &prompt TO hello
This is a prompt: George
1>LIST MEMORY
Name Type Width Contents
PROMPT C 20 "This is a prompt: "
HELLO C 6 George
** Total ** 2 variables... 26 bytes
APPEND [FIELDS <field list> / TEXT <textfile> / OFF ]
Option:
FIELDS <field list> the fields to be edited during APPEND
Option:
OFF rather than generate an APPEND input screen, uses an exiting screen and its Get Table
The APPEND command without either the OFF or TEXT option allows the user
to add records to the selected file. Once the command is given, the screen
shows the new record in full-screen editing mode. For instance, if the
selected file has 201 records, the screen will show record 202 with all fields
filled with blanks; once record 202 is filled in, a blank record 203 is shown.
To exit from APPEND, use <End> (or Ctrl-W) once you have filled in the
fields of the last record desired. If all the fields of the last record are
blank, it will not be appended. To switch from APPEND to EDIT mode, press <PgUp>.
Note that APPEND is actually a special mode of EDIT; the only difference
is that EDIT begins by displaying the current record, while APPEND adds a new
blank record to the data file and displays that. All the editing fields are
the same as in EDIT. See the EDIT command for the complete list of editing
keys.
APPEND updates all index files in use.
Example:
1>USE employee
1>APPEND
Shark goes into screen editing mode and displays record 7 with all
fields blank (the box showing editing keys is displayed when SET MENU is ON):
#1 EMPLOYEE.DBF APPEND Record 7
Page 1
REC: prev <PgUp> next <PgDn> delete ^U PAGE: prev ^K next ^L
FILE: top ^<Home> bottom ^<End> DELETE: char <Del> to end of field ^Y
APPEND MODE: begin ^<PgDn> exit <PgUp> EXIT: with save <End> no save ^Q
NAME...........
FNAME..........
ADDR...........
CITY...........
STATE..........
ZIP............
TEL_NO.........
MARRIED........
SALARY.........
YEAR_EMP.......
DEPT...........
ADD_1..........
Shark offers two options that allow users to format their APPEND
and EDIT screens to their specific requirements.
APPEND OFF suppresses the standard APPEND/EDIT screen and uses an
existing screen and its associated Get Table to accept the input. It is
available only in a program.
Example in a program:
1>USE employee
1>CLS
1>TEXT add_empl
1>APPEND OFF
APPEND TEXT can be used in both conversational mode and programs. The
following combines all the steps in the above example:
Example:
1>USE employee
1>APPEND TEXT add_empl
The input screen can be constructed with either @ GET or TEXT command.
See @ and TEXT commands.
BLANK [append blank record, no full-screen editing]
APPEND BLANK adds a blank record to the selected file, and sets the
current record number to this record. The new record is not shown. The
fields of such a record are usually filled by the user with the REPLACE,
@ GET, or EDIT commands (or variants such as READ, =, TEXT, BROWSE).
APPEND BLANK updates all index files in use. When network is on, the
file header is updated and the file is flushed to disk.
Append data to the selected file from another file.
APPEND FROM <file> [FOR <cond>] [SDF/SDF DELIMITED] [WITH <char>]
<file> the name of the source file
Options:
FOR <cond> select by condition (not with sequential source
files)
SDF sequential source file
SDF DELIMITED sequential source file with delimited fields
SDF DELIMITED WITH <char> sequential source file with fields
delimited with a specific character <char>
This command appends records from the source file <file> to the selected
file. Fields are copied from the source file to the selected file into fields
with the same name. Deleted records are not appended.
If a field in the source file has no matching field in the selected file,
then the field is ignored. If a field in the selected file has no matching
field in the source file, then the field is set to blank.
For character fields, if the field width gets smaller, then the field is
truncated on the right; if the field width gets larger, then the field is
padded with blanks on the right.
For number fields, if a number will not fit in the new field, digits are
lost on the left side of the number. Be careful when moving numbers with
different field widths. If the source field has more decimals than the target
field, decimals on the right will be dropped. If the target field has too few
positions to the left of the decimal point, the value will be lost and the
field filled with a zero followed by asterisks.
The FOR clause is used to append a subset of the file specified. The
condition <cond> is made up of memory variables and the fields of the source
file; permitted only when the source is a data file.
If the structure of two data files is identical in all respects, the
records are appended very quickly; otherwise the APPEND is done one field at a
time. If a field name occurs only in the source file, its values will not be
copied into the target data file. (To change the name of a field, use the
RENAME FIELD command)
APPEND FROM updates all index files in use.
With any of the SDF options, Shark reads the sequential file,
and each line is turned into a new record of the data file appended to the
end. Each line must end with a carriage return (character 13) or a
carriage return-line feed pair (characters 13 and 10). The characters of
the line are placed in the new record one after another from the left.
See the related command: COPY TO ... SDF.
APPEND FROM ... SDF assumes that fields are not trimmed or delimited.
With the SDF DELIMITED option, the line of the sequential file is
regarded as a number of trimmed fields, separated by commas. Strings can, in
addition, be delimited by quotation marks. These fields are placed in the
fields of the data file from the left. The remaining fields of the data
file (if any) are left blank; the remainder of the line (after all the
fields are filled), if any, is ignored.
With the SDF DELIMITED WITH <char> option, the line of the sequential
file is regarded as a number of fields, separated by commas with all
character and logical values surrounded by <char>. Otherwise, this option is
identical to SDF DELIMITED.
The functions to read and write sequential files give much better control
over sequential files.
Important programming note: The APPEND FROM command automatically opens
the FROM file in its internal work area; if the FROM file is already open in
another work area, the compiler will assume it is closed when the APPEND FROM
command is passed during execution, even if that command is in an IF, CASE or
other structure module that is not executed!
Therefore, if there is any reference to the FROM file later in the
program, open the file in the proper work area again immediately after the
APPEND FROM command. If it is not actually needed after the APPEND FROM is
executed (e.g., the program exits after the APPEND FROM), open it with the
COMPILE keyword.
Example:
USE#4 invoices COMPILE
Examples:
Example 1. Appending from a file with the same structure:
1>USE employee
1>COPY STRUCTURE TO empl1
1>USE empl1
1>APPEND FROM employee FOR name < 'Q'
3 APPEND(S)
1>LIST name
1 Marek
2 Balzer
3 Poyner
1>ZAP
1>APPEND FROM employee FOR tel_no < '5'
1 APPEND(S)
Example 2. Appending from a file with a different structure.
Append from EMPLOYEE.DBF:
The selected file is:
Datafile name : EMPL2.DBF
Number of records: 0
Database selected is #1
Field Name Type Width Dec
1 NAME C 15
2 NAME1 C 10
3 ADDR C 5
4 CITY C 25
5 STATE C 2
6 ZIP C 5
7 TEL_NO C 8
8 MARRIED L 1
9 SALARY N 2 0
10 YEAR_EMP N 6 2
11 DEPT C 15
** Record Length ** 95
EMPL2 has a field NAME1 that does not occur in EMPLOYEE; the field FNAME in EMPLOYEE does not occur in EMPL2.
The ADDR field in EMPLOYEE has width 20, in EMPL2 it has width 5.
The CITY field in EMPLOYEE has width 20, in EMPL2 it has width 25.
The SALARY field in EMPLOYEE has width 9, in EMPL2 it has width 2.
The YEAR_EMP field in EMPLOYEE has width 4, in EMPL2 it has width 6; but only 3 characters
are available for the number to the left of the decimal point.
1>USE empl2
1>APPEND FROM employee
6 APPEND(S)
1>EDIT 1
#1 EMPL2.DBF EDIT Record 1
Page 1
REC: prev <PgUp> next <PgDn> delete ^U PAGE: prev ^K next ^L
FILE: top ^<Home> bottom ^<End> DELETE: char <Del> to end of field ^Y
APPEND MODE: begin ^<PgDn> exit <PgUp> EXIT: with save <End> no save ^Q
NAME........... Marek
NAME1..........
ADDR........... 231 R
CITY........... Broomsdale
STATE.......... MD
ZIP............ 02110
TEL_NO......... 566-7012
MARRIED........ y
SALARY......... 0*
YEAR_EMP....... 0*****
DEPT........... Maintenance
Note that ADDR was cut down to 5 characters; CITY was padded by 5 blanks;
SALARY and YEAR_EMP indicate that there was not enough room for the numbers.
Example 3. Appending with the SDF option:
Use the EMPLOYEE file, and append from a sequential file: DATA.TXT. Give
the commands:
1>USE employee
1>APPEND FROM DATA SDF
Let the first line of DATA.TXT be:
Smith Robert 412 River Street
then the first three fields will be correctly placed. Note that the APPEND
with SDF does not check for data correctness (numbers into numeric fields), for
field width, etc. This option can be used if some other program already
formatted the sequential file in absolute conformity with the structure of the
data file.
Example 4. APPEND with the SDF DELIMITED option is easier to use:
Continuing the previous example,
Smith,Robert,412 River Street [or]
'Smith','Robert','412 River Street'
would do. It is important to have the fields delimited with quotation marks if
any field contains a comma. Fields for which no data is supplied should be
represented by commas as "place holders". For instance,
,,'412 River Street'
Example 5. APPEND with the SDF DELIMITED WITH <char> option works the same way:
Continuing the previous example, the command:
APPEND FROM <source> SDF DELIMITED WITH |
working on the following line
|Smith|,|Robert|,|412 River Street|
would append the record correctly. The use of a specified character other than
single quote reduces or eliminates the concern about having commas, apostrophes
and quotation marks in field contents. Fields for which no data is supplied
should be represented by the specified character as "place holders". For
instance,
APPEND TO <num const>
<num const> append record to file in select area <num const>
This command appends the current record from the selected file to the
file numbered <num const>. The same rules apply as in APPEND FROM.
You can also append to the same file by giving its file number.
Examples:
1. Interactive use. There is a Backorder File; you are looking through
it with EDIT or BROWSE to find which orders should be filled. Store
"^Q;APPEND TO 4;BROWSE;^X;" into a function key, and whenever an order to
ship is found, you press the function key. File 4 contains orders to be
shipped.
2. There is a Name-and-Address file, and a label printing program that
prints a whole file, the Label file. The Name-and-Address file is searched by
the above method, and APPEND TO is used to put the addresses to be printed in the Label file.
3. A program segment. File 2 is a backorder file; each record contains 25
fields including Q1 to Q8, the quantities for the eight sizes. This program segment asks how much of each size should be shipped; the end result is that the items to be shipped are appended to the end of File 2, while the original record is changed to reflect the shipped quantities.
SELECT 2
STORE # TO gback
APPEND TO 5
@ y,35 GET q1
@ y,39 GET q2
@ y,43 GET q3
@ y,47 GET q4
@ y,51 GET q5
@ y,55 GET q6
@ y,59 GET q7
@ y,63 GET q8
READ
REPLACE quant WITH q1+q2+q3+q4+q5+q6+q7+q8
SELECT 5
REPLACE q1 WITH q1-q1#2,q2 WITH q2-q2#2,q3 WITH q3-q3#2
REPLACE q4 WITH q4-q4#2,q5 WITH q5-q5#2,q6 WITH q6-q6#2
REPLACE q7 WITH q7-q7#2,q8 WITH q8-q8#2
REPLACE quant WITH q1+q2+q3+q4+q5+q6+q7+q8
REPLACE type WITH 'O'
APPEND TO 2
SELECT 2
GOTO gback
AVERAGE <num exp list> [TO <memvar list> <scope>] [FOR <cond>]
<num exp list> the numeric expressions to average
Options: <memvar list> store the results in these memory variables
<scope> select by scope (default scope: ALL)
FOR <cond> select by condition
The command AVERAGE adds up numeric expressions for selected records of
the selected data file and divides the result with the number of records
summed. Up to 10 expressions can be averaged with one command. Optionally, the results can be stored in numeric memory variables; the expression list and the numeric memory variable list should have the same number of entries. <memvar list> cannot contain numeric mat
Records flagged as DELETED are not averaged.
Memory variables in <memvar list> need not exist; if any variable in the <memvar list> does not exist, this command creates it.
Example:
The average earning of the employees (in the data file EMPLOYEE), and the average year employment began:
1>USE employee
1>AVERAGE salary, year_emp
6 AVERAGE(S)
32502.16 1980.50
Load a binary (assembly-language) program into memory.
BINLOAD <programname>
<programname> a binary file to be executed under Shark; default extension BIN
Assembly-language programs may be copied from disk into a special area
of memory called BINSPACE, which must be set aside with the BINSPACE=
command in your SHARK.SET or SHARKNET.SET file.
When no longer needed, the program can be removed from the BINSPACE with
the BINUNLOAD command, allowing room for another binary program to take its
place. A maximum of eight binary files, with default extension BIN, may be
loaded into memory at one time.
Once loaded into the BINSPACE, the program may be executed with the CALL
command (see the CALL and BINUNLOAD commands).
Reserve space in memory to BINLOAD binary files to be run with the CALL
command.
BINSPACE= (SHARK.SET file only)
<blocks> the number of 1K blocks of RAM to be reserved
for loading BIN files with the BINLOAD command
Assembly-language programs may be copied from disk into a special area
of memory called BINSPACE, which must be set aside with the BINSPACE=
command in your SHARK.SET or SHARKNET.SET file.
The number of 1K blocks, to a maximum of 32, must be specified.
The BINSPACE is allocated above Shark's 64K data space and high
memory, and reduces the amount of DOS memory available to execute commands
with the RUN command.
Once loaded into the BINSPACE with BINLOAD, the program may be executed
with the CALL command (see the CALL, BINLOAD, and BINUNLOAD commands).
Remove a binary (assembly-language) program from memory.
BINUNLOAD <programname>
<programname> a binary file to removed from memory by Shark;
default extension BIN
Assembly-language programs, which are loaded into a special area of
memory called BINSPACE by the BINLOAD command, may be removed from memory by
the BINUNLOAD command when no longer needed.
This allows room for another binary program to take its place. A maximum
of eight binary files, with default extension BIN, may be loaded into memory
at one time.
Care should be taken not to create "holes in memory" by loading and
unloading BIN files indiscriminately. For best performance, users are urged to
load frequently called binary files first and not unload them; then transient
or occasional binary programs can be loaded, called and immediately unloaded
with the BINUNLOAD command.
BOX <line1>,<col1>,<line2>,<col2> [DOUBLE]
<line1>,<col1> the position of the upper-left corner of the box
<line2>,<col2> the position of the lower-right corner of the box
Option:
DOUBLE use double line graphics, the default is single-line
The command BOX draws a box on the screen using the character graphics of
the IBM non-graphics screen. If line1=line2, a horizontal line is drawn. If
<col1>=<col2>, a vertical line is drawn.
<line1>, <line2>, <col1>, <col2> can all be numbers or numeric expressions; any
fractional part will be disregarded.
This command is useful for making fancy menus, boxing messages, and for
partitioning the screen into different viewing areas.
Note that commas are required between the numeric expressions used for
the corner specifications, but a comma is NOT permitted before the DOUBLE.
The WINDOW command can also draw a box, but in addition limits relative
screen output to the boundary of the window and can optionally set window and
box colors. See the WINDOW command.
The BREAK command is used only in programs, in a DO WHILE or a REPEAT
loop to exit at the bottom of the loop. If there are nested loops, the exit
is at the bottom of the innermost loop.
Contrast the command BREAK with the command LOOP which executes the
condition at the top of the DO WHILE loop (see the LOOP command).
Caution: BREAK can be used only within a DO WHILE or REPEAT loop; used
anywhere else will give an error with unpredictable results.
Example:
DO WHILE T
<program segment>
IF custn<>mcust
BREAK
ENDIF
<program segment2>
ENDDO
This program carries out <program segment> and <program segment2>
until CUSTN becomes different from MCUST; then it jumps over <program segment2> and
leaves the loop.
The BROWSE command displays the records (from the current record onward) horizontally, one record per line; it displays as many fields as will fit on a line. A column represents a field.
All fields are available to edit using the BROWSE command. If they don't all fit on your screen, they're visible by simply scrolling the screen left (^K) or right (^L). The scrolling commands are displayed at the top of your screen when you precede the BROWSE command with SET MENU ON.
If you want to see only a few fields, you can use BROWSE FIELDS which will display only the specified fields.
Options (all versions):
FIELDS <field list> specify the fields to be displayed
APPEND add a blank record and start BROWSE in that record
OFF rather than generate a BROWSE input screen, uses an
editing screen and its Get Table
TEXT <textfile> erases the screen, displays the text file, and
then does BROWSE OFF
If the selected file is indexed, the records are displayed in the indexed order.
When the APPEND keyword is used, a blank record is appended to the data
file and BROWSE begins in that record. The user may alternate between normal
BROWSE and APPEND modes with the Ctrl-<PgDn> and <PgUp> editing keys.
All the full-screen editing keys can be used. Note that they have the
same meaning as in EDIT, but sometimes they look different. For instance,
the next record command (<PgDn> or Ctrl-C) gives a new screen in EDIT; in
BROWSE, it simply moves the cursor down one line.
Editing keys:
<left arrow> or Ctrl-S moves the cursor back one character
<right arrow> or Ctrl-D moves the cursor forward one character
Ctrl-<left arrow> moves to the beginning of the field
Ctrl-<right arrow> moves to the end of the field
<Ins> or Ctrl-V puts you in insert mode: what you type gets inserted
(normally, you are in overtype mode: what you type
overtypes the existing text); pressing <Ins> or
Ctrl-V again, puts you back into overtype mode
<Backspace> deletes the character to the left of the cursor
<Del> or Ctrl-G deletes the character on the cursor
Ctrl-Y deletes the rest of the field
<up arrow> or Ctrl-E moves the cursor to the previous field
<dn arrow> or Ctrl-X moves the cursor to the next field
Ctrl-<PgDn> enters APPEND mode; adds a blank record and places the
cursor in that new record. Deactivated by SET SAVE
OFF and SET APPEND OFF; in network mode, deactivated
when data file opened in READ mode
Ctrl-Q quits and does not update the current record
<End> or Ctrl-W quits and updates the current record
<PgUp> or Ctrl-R moves to the previous record; when in APPEND mode, exist
to normal BROWSE mode
<PgDn> or Ctrl-C moves to the next record
Ctrl-L (i.e. ^L) redraws the BROWSE screen with the next screenful of
fields, referred to as the next "page"; if the file
has no more fields, this key is ignored
Ctrl-K (i.e. ^K) redraws the BROWSE screen with the previous screenful of
fields, referred to as the prior "page"; if the
screen is already at the first field, this key is
ignored
Alt-E skips one screenful of record data toward the beginning of
the file and redisplays the BROWSE screen
Alt-X skips one screenful of record data toward the end of the file
and displays the BROWSE screen
Alt-W when cursor is in a memo field, saves the current screen
and opens current memo in read-write mode in the
WRITE editor; on exit from WRITE, the screen is
restored
Alt-R when cursor is in a memo field, saves the current screen
and opens current memo in read-only mode in the WRITE
editor; on exit from WRITE, the screen is restored
BROWSE should be used to edit columns of data in a data file; that is,
to edit a few fields of a number of records. Use the command EDIT to edit
many fields of a single records at the same time.
The option FIELDS <field list> selects the fields to be displayed on
the screen.
SET MENU ON causes a small display at the top of the BROWSE screen,
giving the use of major editing keys.
Examples:
1>USE customer
1>BROWSE
Displays the following:
#1 CUSTOMER.DBF SharkBase BROWSE
Page 1
REC: Prev/PgUp Next PgDn SCROLL: Left/^K Right/^L SCREEN: Up/@E Down/@X
FILE: Top/^Home Bottom ^End MEMO EDIT: @W DELETE: To EOL/^Y Record/^U
APPEND MODE: Begin/^PgDn Exit/PgUp EXIT: With save/End No save/^Q
CUSTNU FIRSTNAME LASTNAME ADDRESS
1 BROS50 Stan Brown 786 Alexander Rd.
2 BURS50 Sid Bursten 876 Main St.
3 GRAG50 George Gratzer 876 Arlington Avenue
4 MELB50 Bernie Melman 9876 Ocean View Parkway
To edit the home and work telephone numbers of the customers, issue the
command:
This is a simpler variation of the BROWSE command that is different enough to be listed as a separate command. Although one can BROWSE fields directly as in "BROWSE NAME, ADDRESS, ETC", the result will be different in look & feel from using the BROWSE OFF command. It can be used in both programs or conversational mode:
SharkBase offers another option that allows the user to format a BROWSE screens to one's specific requirements.
BROWSE OFF suppresses the standard BROWSE screen and uses a much more descriptive formatted display. It is available only in a program.
Example in a program:
*test.prg
USE#1 PRODUCT
SELE 1
CLS
TEXT
Use ↑ or ↓ to scan items - Esc to Quit:
CUST KGS NET COST CHRGE CUST
($Ca) ($US)
@CUST @KGS @COST @US_INVOICE
ENDTEXT
BROWSE OFF
CLOSE ALL
CALL <programname> <argument>
<programname> a binary file to be executed under Shark;
default extension BIN
Option:
<argument> a character memory variable used to transfer data
to and from a BIN program; maximum 254 characters
Assembly-language programs, which have been copied from disk into a
special area of memory called BINSPACE, can be executed internally by Shark.
Running a binary file requires three steps:
1. Allocate memory in the SHARK.SET or SHARKNET.SET file using the BINSPACE=n
where is the number of 1K blocks to allocate to the BINSPACE in memory.
The limit for <n> is 32. Example: BINSPACE=32 allocates 32K.
2. BINLOAD the program. A program need only be loaded once (unless it is
removed with the BINUNLOAD command). Additional requests to BINLOAD the
program will reload it in the same memory space. Up to eight binary programs
may be loaded at once. Example: BINLOAD test. (BIN program may be removed from
memory with the BINUNLOAD command.)
3. CALL the binary program, with an optional argument of up to 254 bytes in
a memory variable. The binary program may modify or replace the contents of
this variable, but may not create or lengthen the contents; when execution is
completed, the variable will have a new value.
See BINLOAD and BINUNLOAD commands.
Example:
The binary program, listed below, merely overwrites the first three
characters of a passed string with the string "SHARK".
dummy='1234567890'
CALL test dummy
? dummy
The current value of dummy would then be printed: "SHARK67890".
Rules for BIN programs:
1. BIN programs are created in assembly language and assembled into an OBJ
file with Microsoft's MASM program or equivalent, linked into a EXE file with
LINK or equivalent, and converted into a BIN file with the DOS utility
EXE2BIN.
2. The following is an example of an assembly-language module that accepts
an input string in a memory-variable passed to the module on the CALL command
line, modifies it, and passes it back in the same memory variable.
;TESTBIN.ASM -- A sample program to illustrate the Shark BINLOAD and CALL facilities
;
;By Bernie Melman
;
;This routine replaces the first five characters in a passed string with the characters SHARK
;
; Assemble with Microsoft Macro Assembler (MASM) version 5.0
; LINK to produce an EXE file (ignore the "No stack" warning)
; EXE2BIN to generate BIN file
;
_prog segment byte
assume cs:_prog
dtest proc far
mov [bx+0], byte ptr 'S'
mov [bx+1], byte ptr 'H'
mov [bx+2], byte ptr 'A'
mov [bx+3], byte ptr 'R'
mov [bx+4], byte ptr 'K'
mov ax,0 ;try changing bp to see if Shark
; can recover regs ok
mov bp,ax ;do it
ret
dtest endp
_prog ends
end
3. No argument is required, but if one is used, it must contain a string
which the program can evaluate in location BX. Shark places a NUL (zero byte) after the string as a terminator; if the program processes characters until a zero is encountered in a byte, the entire string has been processed. Any part of the string following the NUL is ignored by Shark.
4. The maximum length of a Shark string is 254 bytes; therefore, no more
than 254 characters can be communicated to or from a binary program.
5. The BIN program cannot change the memory allocation of the argument
variable. Therefore, it cannot successfully enlarge the argument string. It can shorten the result by terminating the result with a NUL (zero byte).
6. Sufficient space to load binary files must be provided with the BINSPACE=
command in the SHARK.SET or SHARKNET.SET file; and no more than eight binary files may be loaded at any one time. BIN files can be removed from memory with the BINUNLOAD command.
<cond> if the condition is satisfied, the program segment which follows should be executed
CASE is the keyword in the DO CASE program structure. Shark evaluates the condition; if the condition is true, the following program segment is executed. The program segment is terminated by the next CASE, by OTHERWISE, or by ENDCASE. After the execution of the program segment, the program execution continues with the program line after the ENDCASE command. If the condition is false, Shark looks for the next CASE command. If no condition is true, Shark executes the program segment following the OTHERWISE command (if any).
Note that when more than one <cond> is true in a DO CASE structure, only
the program segment for the first is executed.
Leave the current Shark program and start running a new Shark program.
CHAIN <program>
<program> = the name of the Shark program to be run
The command CHAIN is used for executing an Shark program from
within another Shark program or in the interactive mode from the
Shark prompt. The program name <program> should not have an
extension. If there is a compiled program by this name (normal extension
CPL; runtime extension RPL), it will be run. If there is none, Shark will
run the uncompiled program (extension PRG).
CHAIN does a CLEAR first, except that all global memory variables are
preserved and passed to the <program>; to use them, the <program> must have a
GLOBAL command declaring the variables. (See the commands CLEAR and GLOBAL.)
The CHAIN command does not return to the calling program; the program in
memory is replaced by the program it chains to.
CHAIN is the most efficient way for one program to call another. The DO
command calls a subroutine from the disk. DO can often be replaced by PERFORM
or by CHAIN. (See the commands DO and PERFORM).
When you chain from one program to another, Shark executes a
CLEAR command before the start of the program, closing all the data files,
index files, sequential files, and releasing all (but the global) variables.
When you DO one program from another, the subroutine you do inherits the
existing environment -- memory variables, data files, index files, etc. -- and
returns to the DOing program when completed or when a RETURN command is
executed.
CHAIN allows the program name to be a macro.
Example:
TEXT
[1] CUSTOMERS
[2] INVENTORY
[3] ORDERS
ENDTEXT
DO menu; calls a menu program which returns a menu choice as "ans"
DO CASE
CASE ans='1'
CHAIN prog1
CASE ans='2'
CHAIN prog2
CASE ans='3'
CHAIN prog3
ENDCASE
This program segment chains to three different programs, depending on the
value of ANS.
Close all data files and index files, and clear memory variables. Optionally
clear current Get Table or keyboard buffer.
CLEAR [GETS/KEYBOARD]
Options:
GETS clear the Get Table from memory
KEYBOARD delete all characters from the keyboard buffer
buffer
The CLEAR command with no options closes all open data files and index
files, and releases all memory variables, including the matrix variables.
(Sequential files, DOS files, system variables, and in particular, function
keys, are not effected).
Example:
1>a=2
1>name='David Barberr'
1>avco='clear'
1>LIST MEMORY
Name Type Width Contents
A N 8 2
NAME C 13 David Barberr
AVCO C 5 clear
** Total ** 3 Variables used 26 Bytes used
1>CLEAR
1>LIST MEMORY
Name Type Width Contents
** Total ** 0 Variables used 0 Bytes used
CLEAR GETS removes the current Get Table from memory. (See READ and ON
FIELD commands.)
CLEAR KEYBOARD eliminates any characters held in the keyboard buffer.
Normally, characters typed at the keyboard are stored in a special buffer
until Shark is ready to process them, but occasionally the programmer will
want to be sure nothing is in the buffer before executing certain commands.
For example, CLEAR KEYBOARD before executing the MENU( function to ensure
that a key pressed earlier in the program, or even in a previous program, does
not inadvertently trigger a menu selection.
Caution to dBASE programmers: The CLEAR command in dBASE III and later
versions erases the screen, but its function on Shark is radically different.
To clear the screen, use the CLS command.
The CLOSE command closes the selected data file, updates all information
to the disk, and releases the data record buffer space (see Appendix A.1) used
by this file. Any index files attached to the data file are also closed, and
any limits, relations and filters associated with the data file are cleared.
With the option ALL, all data files in use are closed.
This command erases the screen, and is a synonym for ERASE.
If, optionally, two numeric expressions, line1 and line2, are given, it
erases line1 and line2, and all lines between, if any. These expressions
should have values between 0 and 24.
COLOR <attrib>,<line1>,<col1>,<line2>,<col2>[,<fillchar>]
<attrib> numeric expression between 0 and 255, the attribute
<line1>,<col1> the position of the upper-left corner of the box
<line2>,<col2> the position of the lower-right corner of the box
Option:
<fillchar> the ASCII number of the character used to fill the
colored area;
Every character position enclosed
in the area described by the row and column values
is changed to the fill character
For color monitors, both the background and the character are color, identified separately. These attribute bytes are
set for a rectangular area by this command.
The area is given by four numeric expressions -- line1, col1, line2, col2
-- as in the BOX and WINDOW commands. The values must be separated by commas.
Compute ATTRIB by adding up (up to) four numbers (background+
foreground+blink+brightness) from the following tables:
Color Examples:
COLOR 7 (default)
grey on black (7+0+0=7)
COLOR 15
white on black (15+0=15)
COLOR 23
grey on blue (7+16=23)
COLOR 30
yellow on blue (14+16)
COLOR 31
white on blue (15+16=31)
COLOR 36
red on green (4+32=36)
SET COLOR TO 18
Sets DEFAULT to green on blue (2+16=18)
SET COLOR TO 31
Sets DEFAULT to white on blue (15+16=31)
SET COLOR TO 208
Sets DEFAULT to black on pink (0+208=208)
COLOR 112
black on white (0+112=112)
NOTE: The correct syntax within a program is:
COLOR <attrib>,<line1>,<col1>,<line2>,<col2>[,<fillchar>]
When filling an area with a fill character, as is often done in designing
sign-on screens for example, the pattern characters 176 to 178 and the solid
reverse character 219 are especially useful.
See also SET COLOR TO <num exp> and :COLOR=<num exp>, which set the
attribute bytes of the characters displayed. SET COLOR TO 0 turns the command
off: the attribute bytes remain unchanged at the displaced locations. If a
part of the screen already has attributes set by the COLOR command and SET
COLOR TO 0, the newly displayed characters will keep the attributes set by the
COLOR command.
Examples:
1. This snippet
1>COLOR 18,2,0,12,79
sets background blue and foreground green in lines 2 to 12, as in the above table.
2.The various settings can be saved as memvars and recalled:
3. The following Shark program illustrates the use of the COLOR command:
DIM NUM matrix[10]
REPEAT 10 TIMES VARYING i
matrix[i]=MOD(i,7)*16+7
ENDREPEAT
ERASE
REPEAT 10 TIMES VARYING i
COLOR matrix[i], 12-i, i*5, 12+i, 10+i*5
DELAY .5
ENDREPEAT
ERASE
COLOR 23, 0, 0, 24, 79
REPEAT 10 TIMES VARYING i
BOX 1+i, 7+i*3, 23-i, 73-i*3
DELAY .15
ENDREPEAT
REPEAT 10 TIMES VARYING i
COLOR matrix[i], 1+i, 7+i*3, 23-i, 73-i*3
DELAY .5
ENDREPEAT
4. All output to an area of the screen may be made invisible by making
the background and the foreground color the same. For example,
SET COLOR TO 0
COLOR 0,10,0,13,79
turns rows 10 to 13 to black on black. Editing fields for entering passwords
may be so protected.
5. A set of overlapping frames can be created with pattern characters used
as fill:
COLOR 7,10,10,16,70,176 ;a shadow pattern
COLOR 7,09,09,15,69,219 ;a solid pattern
COLOR 7,10,10,14,69,32 ;fill with blanks to create area for text
For more information on setting colors system-wide for all programs at the same time, see the SET COLOR section below.
COMPILE <file> [LINK]
<file> the name of the program file
LINK If SET DO OFF, causes COMPILE to work as if SET DO ON
The COMPILE command may be given in interactive mode, or in a program:
1>COMPILE prog
where PROG is a file with PRG extension, the source program to be compiled.
<file> cannot be a macro, and should not have any extension specified.
If you wish to compile many programs in one step, create a program (call
it, say, PROJECT.PRG) as follows:
COMPILE prog1
COMPILE prog2
COMPILE prog3
Then
1>DO projectj
will compile PROG1, PROG2, PROG3. Such a program should not use any memory
variables, since a CLEAR is executed before every COMPILE command.
SET ECHO ON to have all program lines displayed as they are compiled.
CONTINUE will carry on locating from the current record using the
condition of the last LOCATE command. (See the LOCATE command.)
Note that the standard strings stored in function key F8 combines a
CONTINUE and an EDIT command. Once a LOCATE has been executed (function key
F8), each additional matching record can be edited by pressing F8.
<file> is the name of the file the records are copied to.
Options:
<scope> select records by scope
(default scope: ALL)
FIELDS <field list> copy only selected fields
FOR <cond> select records by condition
SDF copy into sequential file
SDF DELIMITED separate fields by commas
SDF DELIMITED [WITH <char> separate fields by specified character
The command COPY is used for moving records from the selected file to a
new file, <file>; if a file by the name <file> already exists, it will be
overwritten.
If the SDF option is not used, the result will be a data file with the
default extension DBF and the same structure as the source file, unless the
FIELDS keyword is used with a fields list.
When the SDF option is used, the result is a text file, with the default
extension TXT. When the DELIMITED keyword is not used, each record (or its
selected fields) becomes one line in a sequential file; the fields are all
merged.
With the SDF DELIMITED option, the fields are separated by commas, and
strings are enclosed in single quotes ('). A different string delimiter can be
specified with the WITH clause: example WITH |. The specified delimiter
character is not enclosed in quotes.
For a discussion of the SDF option, see the command APPEND FROM.
Examples:
1>USE employee
1>COPY TO empl1
6 COPY(S)
1>USE empl1
1>DELETE 5
1 DELETE(S)
1>COPY TO empl3
5 COPY(S) note: deleted records are not copied
1>RECALL ALL
1>COPY TO empl4 FOR salary <25000 .AND. year_emp>1980
1 COPY(S)
1>COPY TO empl4 FOR salary <25000 SDF DELIMITED WITH |
1 COPY(S)
The result of the last command is a sequential file EMPL4.TXT with one
line as follows:
|Marek|,|Mark|,|231R|,|Broomsdale|,|MD|,|02110|,|566-7012|,|y|,
|11500|,|1984|,|Maintenance|
1>COPY TO empl5 FOR salary <25000 FIELDS name,fname SDF DELIMITED WITH |
1 COPY(S)
The result of the last command is a sequential file EMPL5.TXT with one
line as follows:
<source> = the name of the file to be copied
<dest> = the name of the file after the copy
The COPY FILE command copies a file from the disk to another name
and/or location. As in DOS, a file cannot be copied to itself.
If no extension is given, the extension DBF is assumed. Wildcards are
not permitted.
Important: Do not use COPY FILE on open files, either as source or
destination.
On a network with SET NETWORK ON in SharkBase Network Edition,
attempting to overwrite a file while another user is accessing it will cause
a LOCK error.
Examples:
1>COPY FILE cust a:cust ;data file to floppy
1>COPY FILE cust cust2 ;data file to new name, also DBF
1>COPY FILE cust TO cust.dbk ;file to data backup DBK file
1>COPY FILE edit.prg TO a:edit.prg ;program file to floppy
Create a new file with the structure of the selected file.
COPY STRUCTURE TO <file> [FIELDS <field list>]
Note: <file> = the name of the new file
Option:
FIELDS <field list> copy only these fields
The command COPY STRUCTURE creates a new data file with the same fields
as the selected file, but with no records. The default extension is DBF.
If the FIELDS option is used, the new structure will contain only the
fields listed.
Example:
1>USE employee
1>COPY STRUCTURE TO emp1
1>COPY STRUCTURE TO emp2 FIELDS name,fname,salary,married
1>USE empl2
1>LIST STRUCTURE
Data file: EMPL2.DBF
Number of records: 0
File number: 1
Field Name Type Width Dec
1 NAME C 15
2 FNAME C 10
3 SALARY N 9 2
4 MARRIED L 1
** Record Length ** 36
<scope> select by scope (default scope: ALL)
TO <memvar> store result in memory variable
FOR <cond> select by condition
This command counts the number of records that meet the selection
criteria. Selection is by scope and/or by condition. The result may be
stored in a numeric memory variable; if the variable does not exist, this
command creates it.
Note: <memvar> cannot be a numeric matrix variable.
The result of the command COUNT is displayed if SET TALK ON, as in the
examples below.
Examples:
1>USE employee
1>COUNT
6 COUNT(S)
1>GO TOP
1>COUNT NEXT 3 FOR name <'O'
2 COUNT(S)
1>GO TOP
1>COUNT FOR salary < 50000 TO raise
1>? raise
5.00
The command CREATE is used to make a new data file, <file>; <file> cannot
be a macro. (Do not use COMP for the first four letters of a data or index
file; Shark is unable to open such a file.)
File creation is a special form of full-screen editing. Each field is
represented by four editing fields.
Editing keys:
<Left> or Ctrl-S moves the cursor back one character
<Right> or Ctrl-D moves the cursor forward one character
Ctrl-<Left> moves to the beginning of the editing field
Ctrl-<Right> moves to the end of the editing field
<Ins> or Ctrl-V puts you in insert mode: what you type gets
inserted (normally, you are in overtype mode:
what you type overtypes the existing text);
pressing <Ins> or Ctrl-V again, puts you back
into overtype mode
<BACKSPACE> deletes the character to the left of the cursor
<Del> or Ctrl-G deletes the character on the cursor
Ctrl-Y deletes the rest of the editing field
<Up> or Ctrl-E moves the cursor to the previous editing field
<Dn> or Ctrl-X moves the cursor to the next editing field
Ctrl-Q quits and does not create the file
<End> or Ctrl-W quits and creates the file
Ctrl-K moves back to the top of the previous page
Ctrl-L moves to the top of the next page
Ctrl-N inserts a line for a new field
Ctrl-T deletes the line describing a field
Shark supports four separate type of data files as follows:
Type 0 - The default type; can have up to 256 fields in VP-Info and up to 500 fields in Shark. Allows up to 4 billion records in Shark. In essence, this is an unlimited number of records. The Type 0 file is not compatible with dBASE or early VP-Info versions, whereas the Type 3 variant (see below) is.
Type 1 - Compatible with VP-Info Type 1 files. Type 1 files can have up
to 256 fields in VP-Info, and up to 500 fields in Shark.
Type 2 - Compatible with dBASE II data files, and are limited to 32 fields.
Type 3 - Compatible with dBASE III, dBASE III+ and dBASE IV data files.
Type 3 files can have up to 256 fields in VP-Info and up to 500 fields in Shark (NOTE: dBASE III and dBASE III+ cannot read a data file with more than 128 fields, and dBASE IV cannot read a data file with more than 255 fields). Unlike Type 0 files, Type 3 files are also compatible with all other dBase file types, and also can have up to 4 billion records. Type 3 files are easily the most capable of all existing dBase file types.
Shark includes the earlier Types 1 & 2 file types mostly for compatibility when importing data from older dBase/Clipper files. Data can be read in from older files, and saved in the newer Type 3 file format.
When all the fields are specified using the full-screen input display of the CREATE command, the user is asked to specify which type of file to create; the default is Type 3.
If the data file already exists, you will be given the opportunity to delete it when saving to the new file type. On a network with SET NETWORK ON in Shark Network Edition, attempting to delete a data file while another user is accessing a file with the same name will cause a LOCK error. Wait until the other user exits the file, then continue the save operation.
Example:
1>CREATE custfile
shows the following screen after all the fields have been entered:
Sunday, August 12, 2019
Shark CUSTFILE.DBF
Name Type Width Dec Name Type Width Dec
CUSTNUM C 6 0 BIRTHDAY C 6 0
FIRSTNAME C 15 0 SS_NUM C 9 0
LASTNAME C 20 0 EMPL_NUM C 6 0
ADDRESS C 25 0
CITY C 15 0
STATE C 2 0
ZIP C 9 0
HPHONE C 10 0
WPHONE C 10 0
SPOUSEFNAM C 15 0
SPOUSENAME C 20 0
DEPENDENTS N 2 0
UP/DOWN COLUMN MOVE ROW SAVE STRUCTURE C..Strings
previous. <PgUp> left... ^K insert... ^N update... <End> N..Numbers
next..... <PgDn> right.. ^L delete... ^T nochange. ^Q L..Yes/No
There are up to 12 fields described in each column, with four editing
windows per field. All the standard full-screen editing keys are available in
CREATE, including Ctrl-K and Ctrl-L to move from column to column.
As in BROWSE, <Pg Up> and <Pg Dn> are used to move from line to line,
while <Up> and <Dn> are used to move between editing windows on the same line.
Error checking is done by Shark as you enter the specifications
for the new fields when you leave a line. Here are some of the messages you
may see:
<The first character of a field name must be a letter.
Invalid character in the Name field.
Name fields must be unique.
Types: C-character N-numeric L-logical D-date F-float M-memo.
Field Length must be greater than 0.
Length of character field can not exceed 254.
Length of numeric field can not exceed 20.
Decimals cannot exceed 6.
Decimals too large for length.>
Caution: Be sure you have set FIELDS= in the SHARK.SET or SHARKNET.SET file
to a number large enough to accommodate all the fields in all the data files
you will ever have open at one time, plus the largest number of fields in any
of these files. Default is 320 fields. See FIELDS=.
CURSOR moves the current cursor position to a given point on the screen.
Combined with the ?? command, it gives the user all the absolute positioning
control of the @ SAY command without its restrictions.
CURSOR is usually employed when the "MENU(" function is used, to place the
selection bar correctly.
Print expression or expression list for debugging purposes.
DEBUG <exp list>
<exp list> the expressions to be displayed
This command displays the expressions exactly as does the ? command.
However, if SET DEBUG OFF, the expressions are not displayed. This may save
the programmer the trouble of having to place the DEBUG commands in the
program when debugging, and having to take them out for the regular running
of the program.
See SET DEBUG and Appendix A for more in debugging.
Examples (in a program):
DEBUG number
DEBUG 'current amount: ', amount, ' current balance: ', balance,INKEY()
Note that the INKEY( function causes a pause until any key is pressed.
Suspend execution for a specified number of seconds.
DELAY <num exp>
<num exp> the number of seconds execution is to be delayed
This command suspends program execution for a specified number of seconds. In
some cases, Shark is slow to proceed while clearing its internal memory, and
to proceed immediately will cause an error. In this case, you can insert:
DELAY 3
into the application code, and the application will wait 3 seconds before
continuing.
The delay can be any length from .06 seconds upward. For programmers,
this command can replace complex loops which often have the deficiency of
varying in length depending on what type of computer the program is run on.
The DELETE FILE command deletes a file from the disk. This command
should be used with care because, once deleted, the file cannot be recovered.
If no extension is given, the extension DBF is assumed. Wildcards are not
permitted.
On a network with SET NETWORK ON in Shark Network Edition,
attempting to delete a file while another user is accessing it will cause a
LOCK error.
Use DIM to dimension, or define, an array as a matrix or table. Many variables can be stored under one variable
name <memvar> in the form of a matrix of cells. The number of rows [x1] and columns [x2] are specified,
and the width <width> of the space in each cell is specified. The <width> of the entries of CHAR
arrays defaults to 10 if not specified by the user.
Additional <pages> of tables can be specified by a third input [x3]. SharkBase is limited to 128 variables
in a single program, and arrays are useful for holding large amounts of data under a single variable name.
Matrixes are economical since they're not stored in program memory. Examples would be tables of shipping rates:
DIM CHAR <width> <memvar>[x1,x2,x3],<memvar2>[y1,y2,y3],...
DIM NUM <memvar>[x1,x2,x3],<memvar2>[y1,y2,y3],...
DIM LOG <memvar>[x1,x2,x3],<memvar2>[y1,y2,y3],...
<memvar>, <memvar2> the names of the memory variables that
are defined as an array
Option:
<width> the width of the character variables
Arrays of up to 3 dimensions can be defined for each of the three data
types (CHARACTER, NUMERIC, LOGICAL). The width of the entries of CHAR
arrays is fixed; it is 10 if not specified by the user.
The three dimensions are:
1. Length
2. Width
3. Page number
Note that in the syntax description of DIM, [x1,x2,x3] does not indicate
three 'options'. The 'dimensions' are enclosed in square brackets thus, [ ], and
must be present in a DIM command. x1, x2, x3 are integer numbers greater than
zero, not variables, up to three in number; at least one (length) must be
specified.
Examples:
One dimensional arrays (also called vectors):
DIM CHAR a[11]
(a vector array eleven rows long of CHAR elements, each defaulting to 10
characters in size)
DIM CHAR 20 a[12]
(a vector array of twelve rows of CHAR elements, each 20 characters in size)
DIM CHAR a[12],b[12]
DIM NUM a[21]
DIM LOG a[20]
To fetch customer details from a dbf record:
USE CUSTOMER READ
Set up a one-dimension array named "CUSTDAT" with 9 rows of 50 characters:
N=1
DO WHILE N<10
? "CUSTDAT["+STR(N,2,0)+"]="+CUSTDAT[N]
N=N+1
ENDDO
TABLES:
Two dimensional matrixes (also called tables):
DIM CHAR 25 first[10,20]
(a table array ten rows long, 20 CHAR variables wide, each defaulting to 10
characters in size)
DIM NUM second[5,20]
(The above creates a table array five rows long, 20 NUM variables wide, each defaulting to 8
digits in size)
Three dimensional arrays:
DIM CHAR 10 a[8,10,12]
(The above is a paged table array of 12 pages of 8 rows of 10 CHAR elements)
DIM NUM b[5,5,50]
A mixed definition:
DIM CHAR 12 a[50],b[2,2,11]
Note that the width of both A and B is 12.
Arrays can be used just like all other memory variables except
that only = and STORE can assign values to them. Commands and functions that
create memory variables (such as the command COUNT) or store values in
existing memory variables (such as the command READ or the function READ()
cannot directly use arrays. Arrays cannot be used in a TEXT structure or TEXT
file.
Arrays are stored in high memory (see Appendix A: Shark Compiler). The size
of an array is at most 64K. Use the STATUS command to find out how much high
memory is available. Numeric variables take 8 bytes and logical variables
take 2 bytes of memory for each entry.
Notes: The numbering is from 1. Arrays can be redimensioned in a
program. No part of a DIM command can be a macro.
Examples:
Memory use (in bytes)
1>DIM CHAR 25 name[40] 1,000
1>DIM CHAR fill[10,7] 700
1>DIM NUM b[2,3],total[10,20,5] 48 and 8000
1>DIM LOG abc[5000] 10,000
When using STATUS (or STAT) to inspect the contents of memory, your MATRIX information will appear as in this example. "Width" shows the WIDTH of each row in characters, and DIMENSION shows how many rows of data are in this particular matrix:
Matrix Name Type Width Dimension
CUSTDATA C 40 [43]
DIR <file specification>
Options:
DOS directory name, with optional drive
letter and colon
<file specification> a file name, with optional DOS windcards;
extension is required if present
This command is similar to the DIR command of the DOS operating system:
it displays the list of files on the current disk, together with file size
and date and time created or last modified. Total size of all matching files
is also given along with the bytes remaining on the disk.
If a file name is given in full, the directory will show only that one
file.
A partial listing of the directory can be specified by giving a with wild card characters.
The wild card character ? may be replaced by any single character; the
wild card character * allows any string.
Examples:
1>DIR
SEN_NAME.NDX 1024 10-24-89 1:07a SEN_NUM.NDX 1024 10-24-89 1:07a
SEN_REIN.PRG 757 8-31-89 11:30p SEN_SYST.DBF 53 9-06-89 9:02a
SEN_ZIP.NDX 1024 10-24-89 1:07a SUBDUE.ARC 342658 11-29-89 8:25p
SUBDUE.PRG 2027 11-29-89 7:21p SUBDUE2.PRG 1168 11-29-89 2:29a
349735 bytes in 8 files.
5251072 bytes remaining.
1>dir \*.
LIB DIR 8-04-89 2:42p MAX DIR 8-04-89 1:27p
RELTEST DIR 8-06-89 10:27a SILVERAD DIR 8-07-89 2:32p
SUPER DIR 8-27-89 8:30p TMP DIR 8-04-89 1:28p
V14 DIR 8-23-89 9:22p VAWORK DIR 8-04-89 12:09p
MANUAL DIR 10-30-89 10:53a GEN DIR 11-08-89 12:00a
0 bytes in 10 files.
5251072 bytes remaining.
1>DIR c*.dbf
1>DIR C:\TEMP\*.DBF
CCUST.DBF 522 10-13- 99 5:42p COMMS.DBF 6546 1-06-122 5:11p
COMMS2.DBF 6342 1-07-112 1:00p CUSTJUNK.DBF 7062 9-17-123 3:57p
CUSTOMER.DBF 734 11-29-112 10:43p
21206 bytes in 5 files.
5251072 bytes remaining.
Note on post-Y2K file dates in a Shark file listing: The first epoch (the 20th century to the end of 1999) is numbered as "O" followed by the year, e.g. "099". The zero is dropped, leaving a file date as above of "10-13-99" (Oct 13,1999). The second epoch (the 21st century) is numbered as "1" followed by the year, displaying a file date as above of "11-29-112" (Nov 29,2012). Dates in programs can be easily formatted into your favorite notation using the DATE() function. Example:
Directory listing, following redirection commands of the current FILES
structure.
DIRF <file specification>
Options:
<file specification> a file name, with optional DOS wildcards; an extension is required if present
Shark provides for a FILES structure which lets you specify default drive
letters and/or directories for the various files used by the programs.
Usually file types are grouped into individual directories according to
"skeletons" constructed with wildcards. For example:
The DIRF command is similar to the DIR command, except that the file
specification is compared to the existing FILES structure and, if a match is
made, the redirection in the structure is applied to the file specification,
so that the directory displayed is of the redirected directory, not the
current directory.
It displays a list of all matching files in that directory, together with
file size and date and time created or last modified. Total size of all
matching files is also given along with the bytes remaining on the current
disk. Example:
Note on post-Y2K file dates in a Shark file listing: The first epoch (the 20th century to the end of 1999) is numbered as "O" followed by the year, e.g. "099". The zero is dropped, leaving a file date as above of "10-13-99" (Oct 13,1999). The second epoch (the 21st century) is numbered as "1" followed by the year, displaying a file date as above of "11-29-112" (Nov 29,2012). Dates in programs can be easily formatted into your favorite notation using the DATE() function. Example:
If a file name is given in full, the directory will show only that one
file. The wild card character ? may be replaced by any single character; the
wild card character * allows any string.
<scope> select by scope (default: the current record)
FOR <cond> select by condition
OFF do not display the record number
Options (for listing directories):
LIKE <skeleton> the file specification, with optional wildcards
ON <drive letter> a drive designation, with optional colon
These commands are exactly the same as the LIST commands except for the
following differences: the default scope of the DISPLAY command is the current
record rather than the whole file; the listing is stopped at the bottom of the
screen or current window (about every 20 lines when displaying data file
contents).
See the commands: LIST, LIST FILES, LIST MEMORY, LIST STRUCTURE, and LIST SYSTEM.
In a program, execute a subroutine and, on completion, return to the next
command in the calling program; in Conversational Shark, equivalent to CHAIN.
DO <file>
<file> the name of the program called
In Conversational Shark, DO is the same as CHAIN: it begins the
execution of the compiled program with the name <file> with extension CPL;
otherwise, the uncompiled Shark program <file> will be compiled "on
the fly" and executed, and the compiled file immediately deleted from the disk.
Note that an uncompiled program with a GLOBAL statement cannot be
compiled on the fly if it needs to import variable values, since compile
step includes a CLEAR command.
In a program, the command DO <file> causes the program <file> to be
compiled as a subroutine of the current program. No matter how many
subroutines are called, all will be compiled into the same CPL file as the
calling program. The subroutines become overlays; they are called into
memory when needed.
When the called program executes a RETURN command, the execution resumes
in the current program with the line following the DO command.
The called program can itself DO other subroutines. There can be
subroutines up to 10 levels.
When a subroutine is compiled, all information about data files and
memory variables is coded in the compiled form: the environment is compiled
with the subroutine. Contrast this with the CHAIN command that starts with
a clean slate.
Any time a subroutine is invoked with a DO, the called program becomes an
overlay. If a subroutine is called 5 times, it is compiled 5 times. This, of
course, would make for very bulky programs.
The solution of this problem is very simple. If the subroutine TEST is
not sensitive to the environment (it does not use any fields or memory
variables, or all data files and memory variable names are the same throughout
the calling program), DO it as follows:
...
PERFORM STEST
...
PERFORM STEST
...
*
PROCEDURE STEST
DO TEST
ENDPROCEDURE
This way, TEST becomes a single overlay; it is invoked with PERFORM STEST.
Note: If the same subroutine is called from two places in the same
program, but the two places have different environments, the subroutine cannot
be called from a single procedure, and has to be compiled twice.
For example, assume that TEST carries out some computation on some fields
of the data file HISTORY1 and some memory variables, and then carries out the
same computations on some fields of the data file HISTORY2 and some memory
variables; HISTORY1 is file 2, while HISTORY2 is file 3.
Create two procedures:
PROCEDURE TEST1
SELECT 2
DO TEST
ENDPROCEDURE
*
PROCEDURE TEST2
SELECT 3
DO TEST
ENDPROCEDURE
Now if you need to invoke TEST with the first environment (HISTORY1),
then PERFORM TEST1; otherwise, PERFORM TEST2.
All procedures called by a subroutine must be contained in the PRG file
of the subroutine. A procedure of the calling program cannot have the same
name as a procedure of the subroutine. It is good programming practice not
to have the same name for procedures and subroutines.
Procedures are internally compiled, while subroutines become overlays.
Therefore, procedures are faster in execution. However, there is a limit of
32 procedures in total for one CPL file.
The name of the PRG file used with the DO command cannot be a macro.
Example in Conversational Shark:
DO INVOICE
This will execute INVOICE.CPL if present; otherwise it will compile and
execute INVOICE.PRG, and then delete the CPL file.
The command DO CASE provides for the processing of a number of options
without the use of nested IF commands. It is used in conjunction with the
CASE, OTHERWISE, and ENDCASE commands.
The DO CASE program structure is as follows:
DO CASE
CASE <cond1>
<program segment1>
CASE <cond2>
<program segment2>
...
CASE <condn>
<program segmentn>
OTHERWISE
<program segment>
ENDCASE
where <cond1>, <cond2>, <cond3>,..., <condn> are conditions, <program
segment1>, <program segment2>,..., <program segmentN> and <program segment>
are program segments, that is, any number of Shark program lines.
Execution is as follows:
After the DO CASE is encountered, Shark looks for the first CASE
command and evaluates the condition; if the condition is true, the following
program segment is executed, terminated by the next CASE, by OTHERWISE, or by
ENDCASE. After the execution of the program segment, the program execution
continues with the program line after the ENDCASE command.
If the condition is false, Shark looks for the next CASE
command. If no condition is true, Shark executes the program segment
following the OTHERWISE command (if any).
The OTHERWISE command is optional. Note that if more than one condition
holds, only the segment after the first true condition is executed.
DO case commands can be nested up to 10 levels.
Example:
ans2='P'
@ 22,10 SAY 'E-edit A-add O-order L-look Q-quit ' GET ans2
READ
@ 22,10
DO CASE
CASE UPPER(ans2)='Q'
picked=f
ok=f
LOOP
CASE UPPER(ans2)='E'
@ y,3 GET style
@ y,10 GET color
@ y,14 GET descript
@ y,35 GET q1
@ y,39 GET q2
@ y,43 GET q3
@ y,47 GET q4
@ y,51 GET q5
@ y,55 GET q6
@ y,59 GET q7
@ y,63 GET q8
READ
IF q1+q2+q3+q4+q5+q6+q7+q8=0
DELETE
ENDIF
mpick=t
CASE UPPER(ans2)='O'
REPLACE type WITH 'O'
...
OTHERWISE
IF type='P'
REPLACE type WITH 'O'
ENDIF
picked=f
ENDCASE
When editing with the internal Shark programming editor, Alt-F reformats
the file with all structures properly indented, making it easy to see unbalanced structures. This is a useful tool for programmers since an unbalanced "if-endif" element can totally disrupt a surrounding "do case" construct.
DO WHILE <cond>
<cond> the condition controlling the loop
The command DO WHILE starts the program loop. The structure of the loop
is as follows:
DO WHILE <cond>
<program segment>
ENDDO
Execution is as follows:
After the DO WHILE is encountered, <cond> is evaluated. If <cond> is
true, the Shark program segment (terminated by the ENDDO) is
executed; then <cond> is evaluated again. If <cond> is false, execution
continues with the command following ENDDO.
There can be a DO WHILE command within a DO WHILE command; this is called
nesting. Many level of nesting is permitted; however, too many levels of
nesting will give a compile-time error message:
STACK OVERFLOW
There are two commands that facilitate moving to the top and the bottom
of the loop: LOOP moves the execution to the top of the loop, and BREAK exits
the loop.
There is one more looping command in Shark: REPEAT.
When editing with the internal Shark programming editor, Alt-F reformats
the file with all structures properly indented, making it easy to see
unbalanced structures.
EDIT <recnum> [FIELDS <field list>] [/ OFF] [/ TEXT <textfile>]
Options
<recnum> begin EDIT on record number <recnum>
FIELDS <field list> the fields to be edited
OFF rather than generate an EDIT input screen,
uses an existing screen and its Get Table
TEXT <textfile> erases the screen, displays the text file,
and then does EDIT using the text file
layout and embedded format pictures
The EDIT command without either the OFF or TEXT option allows the user to
view and modify records in the selected file. Once the command is given, the
screen shows the current record (or the record optionally specified with
<recnum>) in full-screen editing mode.
If the FIELDS keyword and a fields list is specified, only those fields
will be available during EDIT:
1>USE employee
1>EDIT FIELDS name, fname, tel_no
otherwise, Shark builds an input screen using
all fields in the current data file.
To exit from EDIT, use <End> (or Ctrl-W) once you have filled in the
fields of the last record desired. To switch from EDIT to APPEND mode, press
Ctrl-<PgDn>, while <PgUp> switches back from APPEND into EDIT mode.
Note that APPEND is actually a special mode of EDIT; the only difference
is that EDIT begins by displaying the current record, while APPEND adds a new
blank record to the data file and displays that. All the editing fields are
the same for both EDIT and APPEND.
EDIT updates all index files in use.
The command SET MENU ON displays the most important editing keys at the
top of the screen, except if OFF or TEXT keyword is used, or if the current
window is less than 80 characters wide or 10 rows deep.
Editing keys:
<Left> or Ctrl-S moves the cursor back one character
<Right> or Ctrl-D moves the cursor forward one character
Ctrl-<Left> moves to the beginning of the field
Ctrl-<Right> moves to the end of the field
<Ins> or Ctrl-V puts you in insert mode: what you type gets inserted
(normally, you are in overtype mode: what you type
overtypes the existing text); pressing <Ins> or
Ctrl-V again, puts you back into overtype mode
deletes the character to the left of the cursor
<Del> or Ctrl-G deletes the character on the cursor
Ctrl-Y deletes the rest of the field
<up arrow> or Ctrl-E moves the cursor to the previous field
<dn arrow> or Ctrl-X moves the cursor to the next field
Ctrl-Q quits and does not update the current record
<End> or Ctrl-W quits and updates the current record
<PgUp> or Ctrl-R edits the previous record; if in APPEND mode, exits to EDIT mode
<PgDn> or Ctrl-C edits the next record
Ctrl-<Pg Dn> Enters APPEND mode; adds a blank record and places the
cursor in that new record
Ctrl-K moves back to the top of the previous page (not with
OFF or TEXT keywords)
Ctrl-L moves to the top of the next page (not with OFF or TEXT keywords)
See also:
EDIT TEXT
EDIT OFF
EXAMPLES
1>USE employee
1>GO 2
1>EDIT
1>EDIT 2
You then see:
#1 EMPLOYEE.DBF EDIT Record 2
Page 1
REC: prev <PgUp> next <PgDn> delete ^U PAGE: prev ^K next ^L
FILE: top ^<Home> bottom ^<End> DELETE: char <Del> to end of field ^Y
APPEND MODE: begin ^<PgDn> exit <PgUp> EXIT: with save <End> no save ^Q
NAME........... Steiner
FNAME.......... Tom
ADDR........... 114 North Pitt St.
CITY........... Lakewood
STATE.......... MD
ZIP............ 02111
TEL_NO......... 596-0017
MARRIED........ y
SALARY......... 35780.00
YEAR_EMP....... 1984
DEPT........... Engineering
To edit only a few fields:
1>USE employee
1>SET MENU ON
1>EDIT FIELDS name, fname, tel_no
This displays:
#1 EMPLOYEE.DBF EDIT Record 7
Page 1
REC: prev <PgUp> next <PgDn> delete ^U PAGE: prev ^K next ^L
FILE: top ^<Home> bottom ^<End> DELETE: char <Del> to end of field ^Y
APPEND MODE: begin ^<PgDn> exit <PgUp> EXIT: with save <End> no save ^Q
NAME........... Marek
FNAME.......... Joe
TEL_NO......... 566-7012
The following examples show EDIT using the TEXT and OFF keywords combined
with an external text file (output from both is identical; OFF allows the
added flexibility of using an internal TEXT structure and/or an ON FIELD
structure):
1>USE employee
1>EDIT TEXT employee
1>USE employee
1>TEXT Employee
1>EDIT OFF
These both display:
EDIT Record 23
NAME........... ARTHUR NEUMANN
ADD_1.......... 4274 MATHERS BLVD. E. UPSON DOWNS
ZIP............ 59768
PHONE.......... 243-5548 (614)
WPHONE......... 643-5657 (614)
EXPERIENCE.....
ACCT STUDENT, WORKED FOR CPA SCARECROW & MOSCOWITZ, BOSTON
COMMENTS
KNOWS INCOME-TAX LAW, WITH SPECIALTIES IN DEPRECIATION AND
TAX SHELTERS; SUGGEST ASSIGN TO CENTRAL OFFICE
See TEXT for a full discussion of the TEXT command. The TEXT used
for the above input screen:
The following demonstration program illustrates use of EDIT OFF.
* EDIToff.prg demonstration program for the EDIT OFF command
* Note the technique used to "wrap" the file...when skipping
* past EOF, program "wraps to top of file, and vice versa,
* using SOUND command to beep (silent if NOEFFECTS in SHARK.SET).
USE customer
ON escape ;what to do when <Esc> is pressed
WINDOW ;cancel small window
CURSOR 23,0 ;put cursor in bottom left corner of screen
CANCEL ;exit program
ENDON
WINDOW
ERASE
*
TEXT
REC: prev <PgUp> next <PgDn> DELETE RECORD <Ctrl>U
FILE: top <Ctrl><Home> bottom <Ctrl><End> DELETE: char <Del> to end of field <Ctrl>Y
APPEND MODE: begin <Ctrl><PgDn> exit <PgUp> EXIT: with save <End> no save <Ctrl>Q
ENDTEXT
WINDOW 8,10,17,69 double
TEXT
.. custnum,!!!-!-99
.. hphone,999-9999 (999)
.. wphone,999-9999 (999)
.. state,!!
Enter customer data:
Number...... @custnum
Name........ %firstname %lastname
Address..... @address
City, State. %city %state
Zip Code.... @zip
Phones (H/W) %hphone %wphone
ENDTEXT
DO WHILE t
EDIT off ;edit using Get Table created with TEXT above
DO CASE
CASE :key=335 ;exit loop when <End> key pressed
BREAK
CASE eof ;skipped past end of file; start again at top
SOUND 11
GOTO top
CASE #<1 ;backed up past beginning of file; start again at bottom
SOUND 11
GOTO bottom
IF eof ;if no undeleted records left in file, get out
BREAK
ENDIF
ENDCASE
ENDDO
WINDOW ;cancel small window
CURSOR 23,0 ;put cursor in bottom left corner of screen
This command is used on a MS-DOS system when the printer is controlled directly by the program. A Windows-based DOS emulator like vDOS or DOSbox will direct print output first to a file (not to a printer), and Windows then processes the file, handing it to the printer or saving the output as a text file.
Thus, in MS-DOS, to start a new page on the printer.
EJECT
This also resets the line and column counters to 0 for the @ commands.
EJECT will attempt to advance the printer whether printing is on or off.
If your printer is on but off-line, your program may appear to hang; either
put the printer on-line, or turn off the printer to release Shark to
continue processing.
When spooling directly to a printer within an MS-DOS system (see SPOOL), the form-feed is placed into the spool file,
and has no effect on the printer until the file is spooled or copied to the DOS printer.
(See the commands @, SET FORMAT TO PRINT, SET PRINT ON, SET EJECT ON/OFF,
SET LENGTH TO, and the functions ROW( and COL(.
Examples:
1>EJECT
In a program, EJECT and SPOOL may be used together to reset the line
counter to zero before printing begins:
All Shark structures are terminated with a matching END command
as shown above. When editing with the internal Shark programming
editor (see WRITE command), Alt-F reformats the file with all structures
properly indented, making it easy to see unbalanced structures.
See DO CASE, DO WHILE, IF, FILES, ON ERROR, ON ESCAPE, ON FIELD, PROCEDURE, REPEAT, and TEXT.
This command erases the screen, and is a synonym for CLS.
If, optionally, two numeric expressions, line1 and line2, are given, it
erases line1 and line2, and all lines between, if any. These expressions
should have values between 0 and 24.
ERASE is the same as the following three commands:
Initiates a program segment module in an ON FIELD structure
FIELD <fieldname>/<fieldnum>
fieldname name of a field in a Get Table
fieldnum number of a field in a Get Table
Each field in a Get Table may have a program segment in a program's ON
FIELD structure which will be executed when the cursor leaves that field, or
optionally when READ or the current record is exited.
All fields in a Get Table are numbered from 1 to 64 in the order in which
they are placed on the screen. When they are painted by @ GET commands, the
order is the same as the order of the @ GET commands in the program; if painted
by TEXT, the order is left-to-right on each line, and then top-to-bottom.
When editing with the internal Shark programming editor (see
WRITE command), Alt-F reformats the file with all structures properly indented,
making it easy to see unbalanced structures.
Set the maximum number of fields available at any one time.
FIELDS= <num>
This setting is used in the SHARK.SET file only
<num> = maximum number of fields allowed in all data files
open at any one time (including Shark's
internal work file); range 128 to 1000
This command sets the maximum number of fields that can be open at any
one time in all open files, taking the place of SET FIELDS TO in earlier
versions of the language (now ignored if encountered by the Shark compiler).
Reserving memory space for a single field requires 16 bytes. Therefore,
increasing the number of fields requested reduces the memory space available
for your program.
The limits for FIELDS= are 128 and 1000, with a default of 320.
Caution: be sure you have set FIELDS to a number large enough to
accommodate all the fields in all the data files you will ever have open at
one time, plus the largest number of fields in any of these files.
Lets you specify drive letters, directories, and/or default file modes for
various files used by Shark.
FILES
FILES [,] [,<mode>]
specification = any file name or "skeleton" (using the * and ?
wildcards
Options:
file direction any legal DOS path, consisting of drive letter
and colon and/or directory path
Option: (Network Edition only):
mode any one of the optional file-opening modes used
in network operation -- LOCK or L, WRITE or W,
READ or R, or SHARE or S. If no mode is set,
or if Network Edition is not in use, the
default mode is LOCK
There is always a FILES Table in memory, set up by the active
FILES ... ENDFILES structure. (This table is empty if there is no active FILES ... ENDFILES structure.) The FILES command allows on-the-fly additions and changes to the FILES Table from either Conversational Shark or programs.
Examples of Form 1 (Only in programs):
FILES
Initiates a FILES ... ENDFILES structure. (See FILES ... ENDFILES.)
(Only in Conversational Shark)
1>FILES
Empties the current FILES Table in memory. (Caution: Do not confuse this
interactive command with the FILES command as used in programs, which always
initiates a FILES ... ENDFILES structure.)
Examples of Form 2 (In either programs or Conversational Shark):
FILES *.dbf,c:\data
Given this command, either in a program or in Conversational Shark,
Shark searches the FILES Table in memory for a matching file
specification. If the file specification *.DBF is found, the new path will
replace the existing path; otherwise, this specification and redirection is
added to the top of the FILES Table.
FILES *.frm
If no <file> direction is specified, file redirection is turned off for files
matching this entry.
Macros are allowed with the FILES command itself, either in programs or
in Conversational Shark. Macros are not permitted in the FILES
... ENDFILES structure.
THE FOLLOWING REFERS TO NETWORK EDITIONS ONLY:
There is a special form of the FILES command which has significance only
in network editions with SET NETWORK ON.
This form adds a <mode> which sets the default file mode for all matching
files. For example,
FILES *.ndx,ndx,READ
sets a file mode of READ for all files with an NDX extension, all of which are
found in the NDX subdirectory of the current directory.
If only <specification> and <mode> are given, the file mode is given to
all files matching that specification wherever they are on the disk. (Note:
two commas are required:
Lets you specify drive letters, directories, and/or default file modes for
various files used by Shark.
FILES
[, ][,<mode>]
ENDFILES
specification any file name or "skeleton" (using the * and ?
wildcards); must not be indented
Options:
file direction any legal DOS path, consisting of drive letter and colon and/or directory path
Option: (Network Edition only):
mode any one of the optional file-opening modes used
in network operation -- LOCK or L, WRITE or W,
READ or R, or SHARE or S. If no mode is set,
or if Network Edition is not in use, the
default mode is LOCK
(Programs and CNF file only) There is always a FILES Table in memory,
set up by the active FILES ... ENDFILES structure. (This table is empty if
there is no active FILES ... ENDFILES structure.)
A FILES ... ENDFILES structure
may be created in a program at any time, and may have any number of specification
lines between the FILES and ENDFILES lines. All existing entries are immediately
cleared and a new FILES Table constructed.
Examples:
*.dbf,c:\data
Causes Shark to add C:\DATA\ to the front of all file names Given this command,
either in an Shark program or in Conversational Shark, Shark searches the FILES
Table in memory for a matching file specification. If the file specification
*.DBF is found, the new path will replace the existing path; otherwise, this
specification and redirection is added to the top of the FILES Table.
FILES *.frm
If no <file direction> is specified, file redirection is turned off for files
matching this entry.
A FILES ... ENDFILES structure is cleared by an empty structure as follows:
FILES
ENDFILES
No macros are permitted in the FILES ... ENDFILES structure, although
macros are allowed with the FILES command itself, either in programs or in
Conversational Shark.
Caution: There is no way to add a comment inside a FILES ... ENDFILES
structure, and no line should be indented. The contents of FILES ... ENDFILES
structures are not affected by the reformatting facility Alt-F of the WRITE
command.
THE FOLLOWING REFERS TO NETWORK EDITIONS ONLY:
There is a special form of FILES...ENDFILES entry which has significance only in network editions with SET NETWORK ON.
This form adds a <mode> which sets the default file mode for all matching files.
For example,
*.ndx,ndx,READ
sets a file mode of READ for all files with an NDX extension, all of which are
found in the NDX subdirectory of the current directory.
If only <specification> and <mode> are given, the file mode is given to
all files matching that specification wherever they are on the disk. (Note: two
commas are required: *.dbf,,WRITE.)
Set the maximum number of files which can be open at one time (requires DOS 3.3 and above)
FILES= <num const>
SHARK.SET file only
<num const> maximum number of files that can be open at any one time; range 21 to 65
Shark has a default maximum of 20 files, unless the operating system
allows more than 20 open files (DOS 3.3 and above) and that a FILES=
statement is present in the CONFIG.SYS file specifying at least <n> files.
As a practical matter, should be in the range 25 to 60; the default
and minimum is 20. Loading more files reduces the maximum size of programs and
reduces speed of such memory-intensive commands as INDEX.
This command may be placed only in the SHARK.SET file, which is executed
by Shark only when first loaded.
Find a record by its index in the selected data file.
FIND <string>
FIND is one of a family of commands that finds a record in an indexed
data file by matching a given string with key values in the index file:
FIND positions the file on the first record in the index matching the FIND
string (no-find positions the file at the top of file)
LAST positions the file on the last record in the index matching the FIND
string (no-find positions the file at the top of file)
NEAREST positions the file on the first record in the index equal to or
greater than the FIND string
SEEK is identical to FIND, except that it searches for the match to the value
of a character expression instead of a string constant (no-find positions
the file at the top of file)
All forms allow a search to be made on a character expression when the
expression is preceded by the macro symbol "&". When the variable var='TAYLOR', all of the following command lines are equivalent:
FIND TAYLOR
FIND &var
SEEK var
In Conversational Shark, just type FIND and the key.
Examples:
1>FIND TAYLOR
1>FIND TOM
If <string> evaluates the same for two records, the second record will
never be found with FIND.
If SET DELETE ON, deleted records will not be found. (See the command SET.)
Numbers must always be treated as strings, even if the key expression is
a numeric field. If NUM is a numeric field of width 2, FIND 1 will not find
1, but FIND &STR(1,2) or FIND &' 1' will work.
For the users' convenience, Conversational Shark converts all
command lines to upper case before execution. So, to find the inventory
numbers that start with AB, type either of the following two commands:
1>FIND AB
1>FIND ab
It follows that
1>FIND Taylor
is understood by Shark as FIND TAYLOR. If you have to find Taylor,
use one of the following forms:
1>FIND &'Taylor'
1>SEEK 'Taylor'
or if name='Taylor' use one of the following forms:
1>FIND &name
1>SEEK name
If FIND is successful, the value of the current record pointer (as shown
by the # and RECNO( functions) is set to the current record number, and the
system variable :NEAR is set to the same number.
If the record is not found, the current record pointer will be set to 0
and the value of :NEAR will be set to the number of the first record in the
index with a key greater than the FIND string; if the index contains no key
greater than the FIND string, :NEAR is set to the bottom of file, and EOF is
set to T (true).
The command for "if not found" is:
IF #=0 or
IF RECNO()=0
In Conversational Shark or in programs with SET TALK ON,
Shark sends the message "NO FIND" when a search is unsuccessful.
Notes:
FIND only works on the active index of the selected data file. If that index
was created with a selection condition, some records may be excluded from the
index. (See INDEX.)
* With SET EXACT ON (see the SET command) only exact matches are found. The
FIND string must be the same length as the index key, and match exactly; if
SET EXACT OFF, the find string may be shorter.
* An index may be made insensitive to capitalization by using the uppercase
functions !( or UPPER( in index creation. This makes use of the FIND command
much easier in Conversational Shark. (See INDEX command.)
* The macro symbol (&) has a wider definition when used with FIND, LAST,
NEAREST and SEEK than with other Shark command. With this group of
commands only, it alerts Shark that the string expression following
is to be evaluated and used as the FIND string. (Macros in all other commands
are limited to fields and variables; dimensioned variables, functions and
concatenation are not allowed.)
The current record of the selected data file is updated on the disk.
This may prevent accidental loss of data due to power failure.
This command updates only the selected data file; it does not update
other data files in use, nor does it safeguard index files, text files, or
DOS files. After a power failure it may be necessary to reindex the selected
file.
<memvar list> the list of memory variables to be declared global
When a program is chained to another program, all the variables that were
defined as global by the GLOBAL command will be passed to the new program.
The GLOBAL command should be the first command of any program. It should
definitely precede any command that involves a memory variable.
There can be only one GLOBAL command in a program. No part of the GLOBAL
command can be a macro. A matrix variable cannot occur in a GLOBAL command.
This is how the command works. In PROG1, you find at the beginning
GLOBAL a,b,c,d
This places in the memory variable table (in the internal Shark data
space, see Appendix A) the variable names: A, B, C, D, in that order. The
variables at this stage have no type or data. When in PROG1, A is
encountered, that will attach type and a pointer to a value to A.
Now PROG1 is chained to PROG2. For PROG2 to receive the values
associated in PROG1 with A, B, C, and D, start with the command:
GLOBAL a,b,c,d
Note, however, that the variable names do not have to match (although type
must).
GLOBAL x,y,c,d
will create the variable X with the contents of A, Y with the contents of B, C
and D retain their names and types.
If you start PROG2 with
GLOBAL a,b,c
then A, B, and C are inherited. If you specify in PROG2 more global
variables, for instance:
GLOBAL a,b,c,d,e
then A, B, C, and D inherit their values and types from PROG1; E becomes the
fifth entry in the memory variable table, with no type or value.
Example:
PROG1 is chained to PROG2 with the command:
CHAIN PROG2
PROG1 wants to pass to PROG2 the values of the following variables:
TRANSACT, PROCESS, CUSTRANGE, TYPE. To accomplish this, start both PROG1 and
PROG2 with the command:
GLOBAL transact,process,custrange,type
Note that when the execution of PROG2 starts, only these four memory
variables exist.
Now suppose that PROG2 is chained to PROG3, and wants to pass to PROG3
the variables TRANSACT, PROCESS, CUSTRANGE, TYPE, and TRDATE.
Then the first command of PROG1 should be:
GLOBAL transact,process,custrange,type
the first command of PROG2 should be:
GLOBAL transact,process,custrange,type,trdate
the first command of PROG3 should be:
GLOBAL transact,process,custrange,type,trdate
Now if PROG3 chains to PROG1, the variables TRANSACT, PROCESS,
CUSTRANGE, TYPE are still passed on, but not TRDATE.
<num const>
GO <num exp>
GOTO <num exp>
GO RECORD <num exp>
GOTO RECORD <num exp>
GO TOP
GOTO TOP
GO BOTTOM
GOTO BOTTOM
<num exp>/<num const> the number to be assigned to #
The commands GO and GOTO are identical. They assign a new value to #,
the current record pointer (see Section 3.4). The data file is repositioned
to the record with the new record number.
The forms:
GO <num exp>
GOTO <num exp>
GO RECORD <num exp>
GOTO RECORD <num exp>
evaluate the numeric expression <num exp>, throw away any fractional part, and assign the value to the current record pointer, #. So, any one of the following
commands:
1>2
1>GO 2
1>GOTO 2
1>GO (2+2)/2
accomplishes the same: the record pointer (#) is set to 2. Note that
1>(2+2)/2
is not allowed. If the selected file is indexed, the index is adjusted accordingly.
GO TOP, GOTO TOP and GO BOTTOM, GOTO BOTTOM are used to position the
current record pointer to the top and bottom of the current data file. If there is no
index in use, the TOP is Record 1, and BOTTOM is the last record. If an index is
in use, then the TOP and BOTTOM are the first and last records according to the index.
Examples:
1>USE employee
1>GOTO 2
1>DISPLAY name
2 Steiner
1>5
1>DISPLAY name
5 Poyner
1>GOTO 1
1>DISPLAY name
1 Marek
1>GO BOTTOM
1>DISPLAY name
6 Wilson
1>GO TOP
1>DISPLAY name
1 Marek
<num const>
GO <num exp>
GOTO <num exp>
GO RECORD <num exp>
GOTO RECORD <num exp>
GO TOP
GOTO TOP
GO BOTTOM
GOTO BOTTOM
<num exp>/<num const> the number to be assigned to #
The commands GO and GOTO are identical. They assign a new value to #,
the current record pointer (see Section 3.4). The data file is repositioned
to the record with the new record number.
The forms:
GO <num exp>
GOTO <num exp>
GO RECORD <num exp>
GOTO RECORD <num exp>
evaluate the numeric expression <num exp>, throw away any fractional part, and
assign the value to the current record pointer, #. So, any one of the
following commands:
1>2
1>GO 2
1>GOTO 2
1>GO (2+2)/2
accomplishes the same: the record pointer (#) is set to 2. Note that
1>(2+2)/2
is not allowed.
If the selected file is indexed, the index is adjusted accordingly.
GO TOP, GOTO TOP and GO BOTTOM, GOTO BOTTOM are used to position the
current record pointer to the top and bottom of the current data file. If
there is no index in use, the TOP is Record 1, and BOTTOM is the last record.
If an index is in use, then the TOP and BOTTOM are the first and last records
according to the index.
Examples:
1>USE employee
1>GOTO 2
1>DISPLAY name
2 Steiner
1>5
1>DISPLAY name
5 Poyner
1>GOTO 1
1>DISPLAY name
1 Marek
1>GO BOTTOM
1>DISPLAY name
6 Wilson
1>GO TOP
1>DISPLAY name
1 Marek
Request for information on Shark commands and related topics.
HELP <string>
Option:
<string> the topic for which help is wanted
The HELP command provides information on the syntax (format rules) and
use of the commands of Shark, and some related topics; it is available
only in Conversational Shark.
The command
1>HELP
displays all the topics for which HELP is available. Type HELP and the topic
description.
The text is contained in the file SHARK.HLP or SHARKNET.HLP, which
must be in the same directory as your version's MSG file, either in the
current directory or on the DOS path. If the file is not found, Shark will
so advise the user.
IF <cond>
<cond> the condition to be evaluated; if true, segment is executed
The IF command allows conditional execution of a program segment. The
general form of this command is:
IF <cond>
<program segment>
ELSE
<ELSE program segment>
ENDIF
This works as follows: When the IF command is found, the condition <cond>
is evaluated. If <cond> is true, the <program segment> (a number of program
lines) is executed. The end of the <program segment> is at the ELSE or at the
ENDIF (whichever comes first).
If <cond> is false, Shark looks for ELSE. If ELSE is found, the <ELSE program
segment> is executed. The execution continues with the program line following the ENDIF.
Note that ELSE is optional.
There can be an IF command within an IF command; this is called nesting.
Many level of nesting is permitted; however, too many levels of nesting will
give a compile-time error message: Stack overflow.
Care should be taken to nest the IF commands properly in a DO WHILE, DO
CASE, or REPEAT program structure. When editing with the internal
Shark programming editor (see WRITE command), Alt-F reformats the
file with all structures properly indented, making it easy to see unbalanced
structures.
IF may not take a macro. Since a line with a macro is not compiled until
runtime, Shark will be unable to handle nesting properly if a macro
is used, resulting in the error message "ENDIF implies IF" or "ELSE implies
IF". For an alternative, see example 4 below.
Examples:
IF count>5
STORE T TO ok
ELSE
IF count<2
STORE F TO ok
ENDIF
? 'ok'
ENDIF
Indenting the program lines helps to show the level of nesting of the
structures. In WRITE, Alt-F indents the lines and capitalizes the command
verbs.
Rewrite:
do while count<5
if name='DAVID'
name=name+fname
enddo
endif
as follows:
DO WHILE count<5
IF name='DAVID'
name=name+fname
ENDIF
ENDDO
Consider the following simple program:
IF num=1
name='DAVID'
ELSE
IF num=2
name='GEORGE'
ELSE
IF num=3
name='TOM'
ELSE
name='UNKNOWN'
ENDIF
ENDIF
ENDIF
This series of nested IF commands (often called "cascading IFs") is
equivalent to the following DO CASE structure:
DO CASE
CASE num=1
name='DAVID'
CASE num=2
name='GEORGE'
CASE num=3
name='TOM'
OTHERWISE
name='UNKNOWN'
ENDCASE
The DO CASE form is much easier to understand.
If <cond> must be a macro, use the following technique instead:
cond2=' ' ;make sure cond2 exists
cond2=&condition ;assign macro to pre-existing variable
IF cond2 ; now IF does not have a macro
...
INDEX ON <str exp> TO <file> [FOR <cond>]
<str exp> the expression providing the key
<file> the name of the index file
Option:
FOR <cond> a selection; only records for which <cond> is true are
included in the index.
The INDEX ON command creates an index file. Index files are used
1. To organize records into an apparent order according to some key for listing, reporting, browsing, editing, etc.
2. To locate records quickly using the FIND, NEAREST, LAST, SEEK, POST and UPDATE commands. Each key must have a separate index file.
3. To make the file act as though it has only records matching a given selection condition and an apparent order according to some key.
The expression <str exp> must evaluate to a string of length at most 60; otherwise, an error message is sent. In addition, the condition itself must be no more than 100 characters long.
(Do not use COMP for the first four letters of an index file; Shark is unable to open such a file.)
Once an index file exists, the data file can be opened for use with the index file; see the USE <data file> INDEX <index file> command. The commands: APPEND, BROWSE, CHANGE, EDIT, REPLACE, READ automatically update the index files in use. There can be many index files in use, all are updated by these commands, but only one index file works with FIND.
If SET TALK ON, INDEX will output progress reports showing the completion of each 100 records, and the total number of records indexed. Large files create temporary files which are then merged to create the ultimate index.
Index expressions should be of type string, but a numeric field (not a numeric expression) can also be used. If you wish to index on a numeric expression, use the key STR(<num exp>) or PIC(<num exp>, <format)) to convert the values to strings. See the functions STR( and PIC(.
It is often helpful to use the upper-case functions, !( or UPPER(, for string expression that mix upper-case letters and lower-case letters, such as names, addresses, and so on. This way, the data is arranged in alphabetical order instead of the ASCII code order which is often confusing. See the functions CHR( and RANK(.
On a network with SET NETWORK ON in Shark Network Edition, attempting to create an index while another user is accessing a file with the same name will cause a LOCK error.
To index on part of a field:
inde on $(CUST,1,3) TO CUSTOMER
To index on two fields:
1>USE employee
1>INDEX ON name+age TO senior
6 RECORDS IN TOTAL INDEXED
To freshen an index, see the REINDEX command, as in this example:
1>USE employee
1>FIND B
27. File is not indexed.
1>INDEX ON name TO employee
6 RECORDS IN TOTAL INDEXED
1>LIST name
2 Balzer
1 Marek
5 Poyner
4 Rayme
3 Steiner
6 Wilson
1>FIND R
1>DISP name
4 Rayme
Note that once the data file is indexed, you can use FIND:
FIND R
1>? name
4 Rayme
For example, a data file has a numeric field AMOUNT containing numbers from 1 to
5,000; to index the file in decreasing order use the key:
STR(10000-amount,6)
Or, a mailing list has 50,000 records, including about 250 from Utah. An
efficient way to select and print a listing of all those in Utah is:
Create an index file, using the index key of an existing SharkBase, VP-Info,
dBASE, or Clipper index file.
INDEX FROM [TO <file>] [FOR <cond>]
<filename> the name of an existing SharkBase, VP-Info, dBASE, Clipper or FoxPro index file
Options:
<file> the name of the SharkBase index file to be created
with the same key expression as ; default
is same as original
FOR <cond> a selection; only records for which <cond> is true are
included in the index
The INDEX FROM command creates an index file using the index expression
of an existing index, even if the existing index is corrupt and even if it's
from any version of VP-Info, dBASE (II, III, or IV), FoxPro or Clipper. If
the same name is given with the TO clause, the original file is overwritten.
Note: All SharkBase and Clipper indexes use the extension NTX, while
all VP-Info, dBase and FoxPro indexes use the extension NDX for their
mutually incompatible indexes. If the FROM file in this command is
specified as an NDX, the TO file may have the same name with an NTX
extension.
When making an index from one created in Clipper, dBASE or FoxPro, be
aware that some expressions from these languages may not be valid in
SharkBase; if an invalid expression is encountered, an error message is
sent. In addition, the <cond> used with the optional FOR clause must be no
more than 100 characters long.
Among the programs shipped with SharkBase is CONV_INV.PRG, which users
with existing applications may use to automate the conversion of indexes.
Whenever SharkBase cannot understand the index expression in the existing
index file, the program prompts for an alternate index expression using
SharkBase functions and operations. When the FROM file is a Clipper NTX
which requires a different expression for its use under SharkBase, both
expressions are retained so Clipper will handle it properly as well.
For users with existing applications which must be retained in dBASE or
FoxPro, you may use the FILES ... ENDFILES structure to point to data and
index files in your dBASE or Clipper directory.
(Note: You may not use COMP for the first four letters of a data or
index file; SharkBase is unable to open such a file with the USE command.)
Once an NTX index file exists, the data file can be opened for use with
it; see the USE <data file> INDEX <index file> command. The commands:
APPEND, BROWSE, CHANGE, EDIT, REPLACE, READ automatically update the index
files in use. There can be many index files in use, all are updated by
these commands, but FIND and its variants works with only one index file at
a time.
If SET TALK ON and in Conversational SharkBase, INDEX FROM will output
reports showing progress and the total number of records indexed. Large
files create temporary files which are then merged to create the ultimate
index.
On a network with SET NETWORK ON in SharkBase Network Edition,
attempting to create an index while another user is accessing a file with
the same name will cause a LOCK error.
Index files are discussed more fully in the Index TO command below.
See also the REINDEX command.
Examples:
1.
1>USE employee
1>INDEX FROM emplnum TO emplnum
6 RECORDS IN TOTAL INDEXED
1>INDEX FROM emplnum
6 RECORDS IN TOTAL INDEXED
1>INDEX FROM emplnum TO empmarry FOR married
4 RECORDS IN TOTAL INDEXED
2.
An application written in dBASE keeps its data and index files in
C:\DATA. The following FILES ... ENDFILES structure is in the CNF files
executed when SharkBase was started:
This allows the DBASENDX.PRG program in the Reference Guide to work
properly, assuming that C:\DATA has no more than a screenful of data files,
and all index files share the first four letters with the data file on
which they are created (an excellent practice).
The command INPUT is used to enter string, numeric, or logical data into
a memory variable, <memvar>. It works the same way as the command ACCEPT
(see ACCEPT), except that string input needs quotation marks.
The type of the <memvar> is determined as follows:
If the data typed in is in quotation marks, it is a string; <memvar> is
of character type.
If the data is a number, <memvar> is of numeric type.
If the data is Y, y, N, n, T, t, F, or f, <memvar> is of logical type.
The optional character string is used as a prompt. A character
expression cannot be used, but a macro is permitted, provided the macro
expression includes quotation marks.
ACCEPT is the preferred way to enter strings, since ACCEPT does not
require quotation marks as delimiters.
<memvar> cannot be a matrix variable.
Examples:
1>INPUT 'Your name: ' TO name
Your name: 'Cathy'
1>INPUT 'Your age: ' TO age
Your age: 32
1>INPUT 'Are you telling the truth? ' TO truth
Are you telling the truth? n
1>LIST MEMO
Name Type Width Contents
NAME C 5 Cathy
AGE N 8 32
TRUTH L 2 .F.
** Total ** 3 variables, 15 bytes
The following illustrates use of a variable instead of a string as the prompt:
1>prompt='"This is a prompt: "'
1>INPUT &prompt TO hello
This is a prompt: 4
1>LIST MEMORY
Name Type Width Contents
PROMPT C 20 "This is a prompt: "
HELLO N 8 4
** Total ** 2 variables... 28 bytes
Initiates a program segment (a KEY procedure) in an ON KEY structure
KEY <number> <caption>
KEY <func_key> <caption>
<number> character number as returned by INKEY( function
<func_key> function key name, in range F1 to F10
Option:
caption string to be printed on line 24 of screen
(space permitting)
This command has two forms, allowing you to specify the key to be
affected either by its number (any character number greater than 255) or by
its name if it is a function key in the range of F1 to F10 inclusive. (F11
and F12 are not supported by SharkBase.)
In either case, all commands that follow until another KEY command or
the ENDON command constitute a procedure that will be executed whenever the
specified key is pressed by the user.
Captions are automatically displayed on line 24 of the screen, and
saved whenever a KEY procedure is executed. When the procedure is complete,
line 24 is redisplayed. If keys are specified by a function key name (F1
through F10), the key name is automatically shown (along with caption). The
spacing between captions is controlled by the number of trailing blanks in
the caption on the KEY command line.
A KEY command with no procedure in an ON KEY structure merely turns off
the named key. This can be extremely useful in deactivating such keys as
Ctrl-<Pg Dn>.
Note: KEY commands in the ON KEY structures cannot be accessed while
SharkBase waits for a keystroke in the MENU( function or the WAIT command.
Examples:
ON KEY
KEY F1 Help
SCREEN 1,2
WINDOW
CLS
TEXT help
WAIT
SCREEN 2,1
KEY F2 Browse
SAVE gets to test
SCREEN 1,2
WINDOW
CLS
TEXT browse
BROWSE off
SCREEN 2,1
RESTORE gets from test
KEY 328
* deactivate up arrow key
KEY 374
* deactivate Ctrl-<Pg Dn> key
KEY F10 Turn Off Keys
ON KEY
ENDON
ENDON
Find the last record matching a given FIND string by its index in the selected
data file.
LAST <string>
<string> the characters to match in the current master index
LAST is one of a family of commands that finds a record in an indexed
data file by matching a given string with key values in the index file:
LAST positions the file on the last record in the index matching the FIND
string (no-find positions the file at the top of file)
FIND positions the file on the first record in the index matching the FIND
string (no-find positions the file at the top of file)
NEAREST positions the file on the first record in the index equal to or
greater than the FIND string
SEEK is identical to FIND, except that it searches for the match to the value
of a character expression instead of a string constant (no-find positions
the file at the top of file)
All forms allow a search to be made on a character expression when the
expression if preceded by the macro symbol "&". When the variable
var='TAYLOR', both of the following command lines are equivalent:
LAST TAYLOR
LAST &var
In Conversational Shark, just type LAST and the key.
Examples:
1>LAST TAYLOR
1>LAST TOM
If SET DELETE ON, deleted records will not be found. (See the command
SET.)
Numbers must always be treated as strings, even if the key expression is a
numeric field. If NUM is a numeric field of width 2, LAST 1 will not find 1,
but LAST &STR(1,2) or LAST &' 1' will be successful.
For the users' convenience, Conversational Shark converts all
command lines to upper case before execution. So, to find the inventory
numbers that start with AB, type either of the following two commands:
1>LAST AB
1>LAST ab
It follows that
1>LAST Taylor
is understood by Shark as LAST TAYLOR. If you have to find Taylor,
either use the string directly with the macro symbol:
1>LAST &'Taylor'
or if name='Taylor' use the following:
1>LAST &name
When LAST is successful, the value of the current record pointer (as
shown by the # and RECNO( functions) is set to the current record number, and
the system variable :NEAR is set to the same number.
If the record is not found, the current record pointer will be set to 0
and the value of :NEAR will be set to the number of the first record in the
index with a key greater than the FIND string; if the index contains no key
greater than the FIND string, :NEAR is set to the bottom of file, and EOF is
set to T (true).
Limit the records available within the current master index to those matching
all or part of the current record's index key.
LIMIT <num exp>
LIMIT <key exp>
LIMIT
Options:
<num exp> the number of characters in the key expression of
the current master index
<key exp> the string expression to match in the current master index
When the current file has an index, Shark can be made to treat the file as though it includes only those records matching the current record's key expression, or some leftmost part of it. When a LIMIT is in effect, no Shark command can access a record outside that limit.
You can most quickly implement this command by using the <num exp> option. The effect is to have Shark set the LIMIT to the leftmost <num exp> characters of the key expression of the master index.
If <num exp> exceeds the length of the key, the effect is identical to setting the LIMIT to the entire index expression. (Maximum index-key length is 60 characters.)
When LIMIT is used with no option, or when <num exp> is zero, the limit is deactivated.
Using LIMIT with <key exp> requires precision in its use to avoid problems that may become extremely serious. Note that the <key exp> expression must be the same as the master index key in use at the moment, or some LEFT( portion of it. Do not attempt to use an expression that does not meet this requirement.
Cautions:
1. FIND should not be used on a file with an active limit. If you choose to do so, be certain that the FIND will be successful. An unsuccessful FIND places the record pointer on record 0 which is by definition not in the index. This causes an error, suspends operation of a program, and turns off the LIMIT.
2. Do not use REPLACE ALL on a field included in an active limit expression, since upon completion, no records will be included in the limit; the result is the same as above.
3. Do not use APPEND, APPEND BLANK or BROWSE APPEND, or enter "append mode" in EDIT, while a limit is active unless you are sure to make all new records equal to the current LIMIT key before leaving the new records. Doing so causes the record pointer to move outside the LIMIT; the result is the same as above.
4. LIMIT should be used with extreme care, and left in effect for the minimum number of commands possible. For instance, if used with BROWSE, LIMIT should be invoked immediately before the BROWSE command and canceled immediately afterward.
5. LIMIT is automatically canceled when the index or the file with which it is active is closed, or the master index is changed with SET INDEX TO n.
6. LIMIT cannot be used with a string constant instead of a <key exp>. The <key exp> can be implemented with a macro.
<scope> select by scope (default scope: ALL)
FOR <cond> select by condition
<exp list> items to list
OFF do not print the record numbers
This command lists the whole selected data file (including the records flagged as DELETED, unless the command SET DELETE ON is given). If the data file is indexed, the listing is by the indexed order.
The records to be listed can be selected by scope and/or by a condition. To list only selected fields of the records, use an expression list.
Examples:
1>USE employee
1>LIST
1 Marek Joe 231 River Drive Broomsdale
2 Balzer Joan 2407 E 38th Street Broomsdale
3 Steiner Tom 114 North Pitt St. Lakewood
4 Rayme Pamela 42368 Wedgewood Dr. Broomsdale
5 Poyner Roger 2757 Regency Road Florington
6 Wilson Robert 16255 Ventura Street Broomsdale
1>GO TOP
1>LIST NEXT 3 fname,name
1 Joe Marek
2 Joan Balzer
3 Tom Steiner
LIST FILES [LIKE <file format>] [ON <drive letter>]
Options:
LIKE <file format> limit to those files matching the file format
specification (default .DBF)
ON <drive letter> show only files on named disk (colon optional)
This command lists the data files, with the default extension DBF, from
the current or other specified disk drive. LIST FILES is provided only for
compatibility with earlier xBase languages; DIR should be used instead.
If the LIKE clause is used, the type of file required can be specified;
the wild cards * and ? (see the DIR command) can be used just as in the
operating system command DIR.
If the ON <drive letter> clause is used, only the files from the drive
specified are listed.
The command LIST MEMORY displays a listing of all non-matrix memory
variables and their contents, and a separate listing of the matrix variables
and their dimensions.
Examples:
1>LIST MEMO
Name Type Width Contents
** Total ** 0 Variables used 0 Bytes used
1>name='David'
1>number=12.78
1>LIST MEMO
Name Type Width Contents
NAME C 5 David
NUMBER N 8 12.78
** Total ** 2 Variables used 13 Bytes used
1>DIM NUM a[12]
1>DIM CHAR table[2,12]
1>LIST MEMO
Name Type Width Contents
NAME C 5 David
NUMBER N 8 12.78
Matrix Name Type Width Dimensions
A N 8 [12]
TABLE C 10 [2,12]
** Total ** 4 Variables used 33 Bytes used
Note that matrix variables are stored in high memory (see Appendix A).
The number of "bytes used" message does not take this into account. To see the
usage of high memory, use the STATUS command. See also DIM.
The command LIST STRUCTURE displays the structure of the selected data file.
Examples:
1>USE employee
1>LIST STRU
Data file: EMPLOYEE.DBF
Number of records: 6
File number: 1
Field Name Type Width Dec
1 NAME C 15
2 FNAME C 10
3 ADDR C 20
4 CITY C 20
5 STATE C 2
6 ZIP C 5
7 TEL_NO C 8
8 MARRIED L 1
9 SALARY N 9 2
10 YEAR_EMP N 4
11 DEPT C 15
** Record Length ** 110
1>USE#5 order
1>LIST#5 STRU
Data file: ORDER.DBF
Number of records: 3
File number: #5
Field Name Type Width Dec
1 ORDER:NO C 6
2 CUST:NO C 6
3 INVEN:NO C 6
4 STYLE:NO C 4
5 COLOR C 2
6 CATEGORY C 2
7 DESCR C 30
8 SLSMAN:NO C 3
9 PRICE N 9 2
10 QTY N 9 2
11 COST N 9 2
12 AMOUNT N 9 2
** Record Length ** 96
List the function key definitions and system variables.
LIST SYSTEM
The command LIST SYSTEM displays the current contents of the 10 function
keys and system variables (see Sections 1.2 and 2.6).
Example:
1>LIST SYST
Name Type Width Contents
:F1 C 5 HELP
:F2 C 10 LIST STRU;
:F3 C 6 WRITE
:F4 C 10 LIST MEMO;
:F5 C 7 BROWSE;
:F6 C 5 STAT;
:F7 C 12 ^wCONT;EDIT;
:F8 C 11 LOCATE FOR
:F9 C 5 FIND
:F10 C 5 EDIT;
:TIME C 8 22:53:43
:DATE C 24 Thursday, March 19, 1991
:VERSION C 4 520
:SERIAL C 10 5208796982
:COMPANY C 34 Sub Rosa Publishing Inc.
:PICTURE C 10 9999999.99
:TITLE C 1
:UNDOC C 1 2003.03.14 <-- you can store application data in system
:KEY N 8 13
:AVAIL N 8 0
:FIELD N 8 0
:ERROR N 8 0
:MESSAGE C 1
:RETRY N 8 25
:USER N 8 1
:DIR C 11 C:\SHARK
:NEAR N 8 0
:COLOR N 8 48
** Total ** 28 variables... 244 bytes
To set the system variables to suit your needs, use your CNF file file
as appropriate, see Section 1.
LOCATE <scope> FOR <cond>
Option: <scope> search only within the scope (default scope: ALL)
The command LOCATE searches for records by a condition. This command
works quicker if there is no index file. If a match is found with SET TALK ON, Record n
informs the user of the record number (n).
The LOCATE command line cannot be more than 128 characters. The keyword
FOR cannot be in a macro.
The CONTINUE command continues the search after the current record using
the condition of the last LOCATE.
Normally, the selected file should not be indexed. The command LOCATE
executes several time slower in an indexed file. If an index file is in
use, the search will take place in the indexed order.
The preferred way to locate a record in an indexed file is with the FIND command.
With the default scope ALL, LOCATE starts at the top of file and scans
either until the first record matching the condition is found, or the end of
file is reached; an unsuccessful LOCATE sets the EOF flag to T (true).
You can use an alternate scope, NEXT or WHILE, to have LOCATE begin with
the record after the current record. Unlike all other uses of scope in Shark, however,
LOCATE does not require anything other than the keyword to make the scope
act in this way.
Example:
LOCATE NEXT FOR cust>'M' ;looks for a matching record anywhere in the
file, starting with the next record.
LOCATE WHILE FOR cust>'M' ;works exactly the same.
Examples:
1>USE employee
1>LIST name, salary
1 Marek
2 Balzer
3 Steiner
4 Rayme
5 Poyner
6 Wilson
1>GO TOP
1>LOCATE FOR salary > 20000
Record 1
1>? name, salary
Marek 25800.00
1>CONT
Record 3
1>? name, salary
Steiner 35780.00
1>CONT
Record 4
1>? name, salary
Rayme 79110.00
1>CONT
Record 6
1>? name, salary
Wilson 33870.00
1>CONT
1>? EOF
T
In Shark Network Edition, this command locks the current record to all
other users so that you can safely make changes to a record key while the
file is open in SHARE mode.
If SET NETWORK OFF, LOCK is ignored.
In the single-user edition, LOCK has no effect, but is accepted
for program compatibility across versions.
See the SET NETWORK command on how to get into the multi-user mode. See
also UNLOCK.
In Shark Network Edition this command locks all indexes attached to the
current data file so that you can safely make changes to a record's key while
the file is open in SHARE mode.
Like the LOCK command (which locks records in a data file), and SET LOCK
ON (which automatically locks each record as it is used), LOCK INDEX should be
used only when files are opened in SHARE mode on a network.
LOCK INDEXES allows only one person to change the index files, though any
number of others can access the indexes if they open the data file in either
SHARE or READ mode. UNLOCK INDEXES will unlock the all the index file.
See the SET NETWORK command on how to get into the multi-user mode. See
also UNLOCK INDEXES.
In the single-user edition, LOCK INDEXES has no effect, but is
accepted for program compatibility across versions.
When the LOOP command is found in a DO WHILE <cond> loop, program
execution jumps to the top of loop (evaluates the condition). See also DO
WHILE and BREAK.
Example:
USE cust
USE#2 trans INDEX trans1
DO WHILE T
ACCEPT 'Enter the customer number ' TO mcust
IF TRIM(mcust)=' '
BREAK
ENDIF
FIND#2 &mcust
IF RECNO(2)=0
LOOP
ENDIF
SELECT 2
DO WHILE custn=mcust .AND. .NOT. EOF
? date,descript,amount
SKIP
ENDDO
SELECT 1
ENDDO
Scan selected records for the highest values of one or more numeric
expressions.
MAX <scope> <num exp list> [TO <memvar list>] [FOR <cond> ] <num exp list>
the numeric expressions to sum
Options:
<scope> select by scope (default scope: ALL)
<memvar list> store the results in these memory variables
FOR <cond> select by condition
The command MAX evaluates numeric expressions for selected records of the
selected data file, determining the highest value encountered for each
expression. Up to 10 expressions can be scanned with one command.
Optionally, the results can be stored in numeric memory variables; the
expression list and the numeric memory variable list must have the same number
of entries. <memvar list> cannot contain numeric matrix variables.
Records flagged as DELETED are not scanned. See also MIN, SUM, COUNT and
AVERAGE.
Example:
The highest hourly wages, the highest annual earnings, and the highest
number of hours worked in the year:
1>USE employee
1>MAX hourly,ydt_earned,ytd_hours
6 RECORDS CHECKED
32.57, 51579.65, 2632.65
1>SET TALK OFF
1>MAX hourly,ydt_earned,ytd_hours TO maxrate,maxearned,maxhours
1>? maxrate,maxearned,maxhours
32.57 51579.65 2632.65
MEMORY=<blocks> (SHARK.SET file only)
<blocks> number of 1K blocks of memory to reserve for high memory
The command MEMORY= in the SHARK.SET file allocates a specific amount of
high memory for use of Shark rather than the default of 128K.
The number of blocks can be as little as 32, although Shark
can be expected to be sluggish in performance at that level, and there will be
little space available for matrixes and background screens.
High memory is used for a wide variety of tasks in Shark, including storage of
index blocks, print spooling, background screen storage, matrixes, table
storage, program swap area, etc. Requesting 32 to 127 blocks
decreases the space available for these tasks, but increases the space
available for storing binary files to be executed with the CALL command (see
BINSPACE= command) and for running programs in DOS (see RUN command).
If large amounts of memory are required for matrixes, you may need to
request more than 128 blocks. Requesting more memory than available will not
cause an error; Shark always takes all the memory there is or
the amount requested, whichever is less.
See Appendix A for more about organization of memory.
Scan selected records for the lowest values of one or more numeric expressions.
MIN <scope> <num exp list> [TO <memvar list>][FOR <cond> ]
<num exp list> the numeric expressions to sum
Options:<scope> select by scope (default scope: ALL)
<memvar list> store the results in these memory variables
FOR <cond> select by condition
The command MIN evaluates numeric expressions for selected records of the
selected data file, determining the lowest value encountered for each
expression. Up to 10 expressions can be scanned with one command.
Optionally, the results can be stored in numeric memory variables; the
expression list and the numeric memory variable list must have the same number
of entries. <memvar list> cannot contain numeric matrix variables.
Records flagged as DELETED are not scanned. See also MAX, SUM, COUNT and
AVERAGE.
Example:
Compute the lowest hourly wages, the lowest annual earnings, and the lowest number
of hours worked in the year:
1>USE employee
1>MIN hourly,ydt_earned,ytd_hours
6 RECORDS CHECKED
5.57, 1579.65, 232.65
1>SET TALK OFF
1>MIN hourly,ydt_earned,ytd_hours TO maxrate,maxearned,maxhours
1>? maxrate,maxearned,maxhours
5.57, 1579.65, 232.65
The command MODIFY (or MODIFY STRUCTURE) modifies the structure of the
selected file. The display is the same as in CREATE, the keys retain their
meaning. Each field is represented by four editing fields.
All records in the file will be automatically appended. If any types or
names are changed, the contents of the affected fields will be lost.
Restrictions:
Do not use MODIFY in a program
Do not use MODIFY merely to change the name of a field. Fields may be
renamed instantaneously with the RENAME FIELD command.
Editing keys:
<Left> or Ctrl-S Moves the cursor back one character
<Right> or Ctrl-D Moves the cursor forward one character
<Ctrl-Left> Moves to the beginning of the editing field
<Ctrl-Right> Moves to the end of the editing field
<Ins> or Ctrl-V Puts you in insert mode: what you type gets inserted
(normally, you are in overtype mode: what you type
overtypes the existing text); pressing <Ins> or
Ctrl-V again, puts you back into overtype mode
<BACKSPACE> Deletes the character to the left of the cursor
<Del> or Ctrl-G Deletes the character on the cursor
Ctrl-Y Deletes the rest of the editing field
<Up> or Ctrl-E Moves the cursor to the previous editing field
<Dn> or Ctrl-X Moves the cursor to the next editing field
Ctrl-Q Quits and does not modify the file
<End> or Ctrl-W Quits and modifies the file
Ctrl-K Moves back to the top of the previous page
Ctrl-L Moves to the top of the next page
Ctrl-N Inserts a new line
Ctrl-T Deletes the line
To be safe, make a back up copy of the file to be modified before the
MODIFY command is given. After the MODIFY, the new fields will be blank, the
deleted fields disappear. All deleted records will also be lost.
If or Ctrl-Q is pressed at any time in MODIFY before the appending
of records is completed, MODIFY will be aborted and the data file closed.No
changes already made will be saved.
Changing the characteristics of a field (say the width) effects the field
the same way as in APPEND FROM.
On a network with SET NETWORK ON in Shark Network Edition,
attempting to MODIFY a file while another user is accessing it will cause a
LOCK error. Always open a file in LOCK mode before modifying it.
Example:
Change the FNAME field in EMPLOYEE.DBF from 10 characters to 15 characters.
1>USE employee
1>COPY employee to temp making a backup copy
1>MODIFY
Thursday, March 22, 1990
Shark Modify EMPLOYEE.DBF
Name Type Width Dec Name Type Width Dec
NAME C 15 0
FNAME C 10 0
ADDR C 20 0
CITY C 20 0
STATE C 2 0
ZIP C 6 0
TEL_NO C 10 0
MARRIED L 1 0
SALARY N 9 2
YEAR_EMP N 4 0
DEPT C 15 0
UP/DOWN COLUMN MOVE ROW SAVE STRUCTURE C..Strings
previous. <PgUp> left... ^K insert... ^N update... <End> N..Numbers
next..... <PgDn> right.. ^L delete... ^T nochange. ^Q L..Yes/No
Now change 10 to 15:
Thursday, March 22, 1990
Shark Modify EMPLOYEE.DBF
Name Type Width Dec Name Type Width Dec
NAME C 15 0
FNAME C 15 0
ADDR C 20 0
CITY C 20 0
STATE C 2 0
ZIP C 6 0
TEL_NO C 10 0
MARRIED L 1 0
SALARY N 9 2
YEAR_EMP N 4 0
DEPT C 15 0
UP/DOWN COLUMN MOVE ROW SAVE STRUCTURE C..Strings
previous. <PgUp> left... ^K insert... ^N update... <End> N..Numbers
next..... <PgDn> right.. ^L delete... ^T nochange. ^Q L..Yes/No
Find the first record whose index key is equal to or greater than a given FIND
string by its index in the selected data file.
NEAREST <string>
<string> the characters to match in the current master index
NEAREST is one of a family of commands that finds a record in an indexed
data file by matching a given string with key values in the index file:
NEAREST positions the file on the first record in the index equal to or
greater than the FIND string. If no key in the index is equal to or
greater than the FIND string, the current record pointer and :NEAR are
set to the bottom of file, and EOF is set to T (true)
FIND positions the file on the first record in the index matching the LAST
string (no-find positions the file at the top of file)
LAST positions the file on the last record in the index matching the FIND
string (no-find positions the file at the top of file)
SEEK is identical to FIND, except that it searches for the match to the
value of a character expression instead of a string constant (no-find
positions the file at the top of file)
All forms allow a search to be made on a character expression when the
expression if preceded by the macro symbol "&". When the variable
var='TAYLOR', both of the following command lines are equivalent:
NEAREST TAYLOR
NEAREST &var
In Conversational Shark, just type NEAREST and the key.
Examples:
1>NEAREST TAYLOR
1>NEAREST TOM
If SET DELETE ON, deleted records will not be found. (See the command SET.)
Numbers must always be treated as strings, even if the key expression is
a numeric field. If NUM is a numeric field of width 2, NEAREST 1 will not
find 1, but NEAREST &STR(1,2) or NEAREST &' 1' will be successful.
For the users' convenience, Conversational Shark converts all
command lines to upper case before execution. So, to find the inventory
numbers that start with AB, type either of the following two commands:
1>NEAREST AB
1>NEAREST ab
It follows that
1>NEAREST Taylor
is understood by Shark as NEAREST TAYLOR. If you have to find
Taylor, either use the string directly with the macro symbol:
1>NEAREST &'Taylor'
or if name='Taylor' use the following:
1>NEAREST &name
By definition, NEAREST is always successful (unless the file is empty).
If an exact match is found, both the current record pointer (as shown by the #
and RECNO( functions) and the system variable :NEAR are set to the first
matching record.
If an exact match is not found, the current record pointer and the value
of :NEAR will be set to the number of the first record in the index with a key
greater than the FIND string; if the index contains no key greater than the
FIND string, the current record pointer and :NEAR are set to the bottom of
file, and EOF is set to T (true).
To position the record pointer to the record equal to or immediately less
than the FIND string, do the following:
NEAREST &findstring
IF keyexpr<findstring
SKIP -1
ENDIF
This command in the SHARK.SET file suppresses the special screen and
sound effects used for the sign-on screen, as well as for the SCREEN, SOUND,
and RING commands.
This is the same as the command * (see *). Any Shark command
line that starts with NOTE is ignored by Shark.
Notes can also be added to any command line with a semicolon(;)
Since notes are not compiled, they have no effect on the execution of
programs. Comments should be used liberally to help you (and possibly others)
understand your programs later. If programs are very large, however, some
comments may have to be shortened or eliminated to allow the compiler and/or
WRITE to load them properly.
Examples:
NOTE This is the start of the main menu display (Shark ignores this line)
IF cust=mcust ;the start of the processing for the current customer (Shark ignores text after ";")
Do not write changes made to the current record back to the disk.
NOUPDATE
Whenever a change is made to the current record in an open data file,
leaving the record or closing the file causes the changes to be written to the
disk. This automatic action can be overridden by the NOUPDATE command, which
tells Shark to ignore the changes.
This is essentially what happens when Ctrl-Q is pressed during EDIT and BROWSE.
Example in a program:
Suppose full-screen editing is in effect with the READ command, and you
want to ignore any changes made to the record if the user presses Ctrl-Q or an
erroneous key like Ctrl-T and Ctrl-U. Using the :KEY system variable that
stores the key number of the key pressed to terminate the last process:
DO WHILE t
READ
IF :key=17 .OR. :key=20 .OR. :key=21 ;key numbers of ^Q, ^T and ^U
NOUPDATE
LOOP
ENDIF
DO CASE
CASE :key=... ;check a key
... ;action based on :key value
ENDCASE
ENDDO
Tell Shark what to do if an error is encountered during program execution.
ON ERROR
The command ON ERROR starts the ON ERROR/ENDON structure. There can be
any program segment between ON ERROR and ENDON, except that no data files
should be opened or closed, and no new variables created or released. This
program segment is executed whenever the program encounters an error.
The ON ERROR is in effect in the program from the point where it is
located. It remains in effect until another ON ERROR structure is found, or
until execution leaves the current module, either on a RETURN or DO command.
ON ERROR has no effect within a subroutine called by the DO command; if
the subroutine needs the same ON ERROR structure, repeat the structure in the
subroutine.
Since the ON ERROR structure has no effect after returning to a calling
program from a subroutine, you should consider putting this structure
(possibly together with the ON ESCAPE structure) into a procedure that is
performed at the beginning of the program, and after every DO command.
ON ERROR should normally include a CHAIN, CANCEL, or QUIT command
to allow execution to leave the current program. Do not RETURN from a main
routine.
When editing with the internal Shark programming editor (see
WRITE command), Alt-F reformats the file with all structures properly
indented, making it easy to see unbalanced structures.
Example:
ON ERROR
? 'Aborting program execution because Error',:error,'encountered'
? :message ; :error contains the latest error number, and
? ; :message contains its meaning
RING
CANCEL
ENDON
The command ON ESCAPE starts the ON ESCAPE/ENDON structure. There can be
any program segment between ON ESCAPE and ENDON, except that no data files
should be opened or closed, and no new variables created or released. This
program segment is executed whenever the user presses .
The ON ESCAPE is in effect in the program from the point where it is
located. It remains in effect until another ON ESCAPE structure is found, or
until execution leaves the current module, either on a RETURN or DO command.
ON ESCAPE has no effect within a subroutine called by the DO command; if
the subroutine needs the same ON ESCAPE structure, repeat the structure in the
subroutine.
Since the ON ESCAPE structure has no effect after returning to a calling
program from a subroutine, you should consider putting this structure
(possibly together with the ON ERROR structure) into a procedure that is
performed at the beginning of the program, and after every DO command.
ON ESCAPE should normally include a CHAIN, CANCEL, or QUIT command to
allow execution to leave the current program.
When editing with the internal Shark programming editor (see
WRITE command), Alt-F reformats the file with all structures properly
indented, making it easy to see unbalanced structures.
Example:
ON ESCAPE
? 'Aborting program execution because key pressed.'
RING
CANCEL
ENDON
Control the processing of data input during execution of the full-screen
editing commands: READ, EDIT, APPEND, and BROWSE.
ON FIELD
The command ON FIELD starts the ON FIELD/ENDON structure. There can be
any program segment between ON FIELD and ENDON, except that no data files
should be opened or closed, and no new variables created or released.
Inside the ON FIELD structure is a number of FIELD statements, each
telling Shark what to do when the user leaves a specific field,
enters a new record, or leaves the record. There can be one FIELD statement
for each input field in a Get Table, in addition to the two FIELD statements
controlling entry and exit from READ (or the current record in EDIT and
BROWSE).
(Note that the term editing field does not refer in this case only to
fields, but to the on-screen area where you're expected to enter or edit data;
an editing field may be a "window into" a data-file field, a memory variable,
or a system variable. An element in a matrix cannot be accessed with the
full-screen editing commands, so cannot be tied to an editing field.)
ON FIELD is used primarily for the following three functions:
1. To verify data input. If the value entered does not meet a condition, the
cursor can be directed back to the current field with the
:FIELD=FIELD(fieldname) command.
2. To do data lookups from the selected or other files. For example, suppose
the first field in the screen is mcust; when a customer number is entered FIND
&mcust is executed and the proper record found in the current file. If an
invoice number is entered, data can be retrieved from the proper invoice with
a FIND into another indexed file.
3. To change the order in which the cursor goes to the next field. For
example, a TEXT can have three columns of input figures; Shark normally
goes left to right on each line, and then top to bottom, but appropriate
use of the :FIELD system variable can direct the cursor to do one column at
a time.
ON FIELD exception handling only works with memvars (eg: 'MC') and
not with dbf fields (eg: 'CA1'). The data handling works OK, but the
ON FIELD doesn't catch the exceptions.
The form of the ON FIELD structure:
ON FIELD
FIELD 0
<initiating command sequence> ;The FIELD 0 code is executed when READ is
entered or whenever a new record is entered in BROWSE OFF, EDIT OFF, etc.
Field 1
<command sequence 1> ;The FIELD 1 code is executed when the
cursor leaves the first editing field (and again when the current record is
exited provided SET EXECUTE ON).
Field 2
<command sequence 2> ;The FIELD 2 code is executed when the
cursor leaves the second editing field (and again when the current record is
exited provided SET EXECUTE ON).
Field <n>
<command sequence <n> ;The FIELD <n> code is executed when the
cursor leaves the nth (final) editing field (and again when the current record
is exited provided SET EXECUTE ON).
Field 65
<command sequence 65> ;The FIELD 1 code is executed when the
cursor leaves READ or when the current record is exited in BROWSE OFF, EDIT
OFF, etc.
ENDON
Fields 1 through 64 may be referenced either by number of by the name of
the field or variable being edited. Names are usually more understandable and
always more flexible, since a field can later be added or omitted without
having to renumber the entire ON FIELD structure.
You can force Shark to execute the entire set of FIELD commands
from 1 through 64 by SET EXECUTION ON. The default is OFF. Do not use ON if
there is a possibility of creating an endless loop which doesn't allow a
graceful exit.
The Get Table is usually canceled as soon as execution leaves the current
program, even when just executing a subroutine with the DO command. You can
SET GET ON to maintain the Get Table, provided it is not cleared in the sub-
program. (A Get Table is cleared either with the CLEAR GETS command, or by
creating a GET with TEXT or an @ GET command after a READ command.) The
default is SET GET OFF.
When editing with the internal Shark programming editor (see
WRITE command), Alt-F reformats the file with all structures properly indented,
making it easy to see unbalanced structures.
Example:
date1=date(1)
CLS
TEXT
.. date1,99/99/99
Enter game date (must be a Wednesday)
@date1
ENDTEXT
ON FIELD
FIELD date1
IF DATE(FULL,date1)<>"W"
:field=FIELD(date1) ;use FIELD( to get correct number
ENDIF
ENDON
DO WHILE t
READ
<command modules>
ENDDO
Note the placement of the three elements: the TEXT with its input macro
and the ON FIELD come before the READ or other full-screen editing command,
and are usually outside the DO WHILE loop unless the screen is rewritten
elsewhere inside the loop.
When filling in forms in a text screen, Shark moves from left to right. If you wish to
fill in fields from top to bottom, you will need to use the field
names/numbers to cause Shark to move to the field BELOW rather than the
field to the right:
ON FIELD
FIELD 0
@ 0,0 say "Edit Image of Customer Record "
@ 0,52 SAY "Screen 1 File:"+mrnum
color 31,24,0,24,78
:FIELD=1
FIELD 1
:FIELD=3 <-- This field is BELOW Field 1 in a 2 column screen.
FIELD 3
:FIELD=5
FIELD 5
:FIELD=6
FIELD 6
:FIELD=7
FIELD 7
:FIELD=2
FIELD 2
:FIELD=4
FIELD 4
:FIELD=8
FIELD 8
:FIELD=9
FIELD 9
ENDON
Begin ON KEY structure, which contains one or more KEY modules defining
actions to execute when specific keys are pressed.
ON KEY
The ON KEY structure, like the other ON structures in SharkBase, is not
executed "in line'' as are other SharkBase commands; it is executed only
when a specific condition is encountered.
In the case of the ON KEY structure, every time a key is pressed its
key number is checked; if it is in the range 256 through 511, the ON KEY
structure is checked to see if a special program segment has been specified
for that key number and, if it has been, the program segment is executed.
Each KEY command introduced a program segment that is essentially a
complete procedure, and like other procedures in SharkBase require a stable
environment. Any program segments can be between ON KEY and ENDON, except
that no data files may be opened or closed, and no new variables created or
released. The data files open when the ON KEY structure is executed must
remain use as long as the ON KEY is in effect.
In addition, the current work area for the program segments in the ON
KEY structure will be the same as the work area in effect when the structure
is first encountered. Therefore, an ON KEY program segment cannot be used
to change the work area
There can be a maximum of 32 KEY statements in an ON KEY structure.
All existing Key procedures are terminated by CHAINing or exiting from a
program, or by encountering another ON KEY/ENDON structure.
ON KEY structures can be contained inside ON KEY structures. An ON KEY
structure with no KEY commands is used to deactivate all KEY procedures.
Note: KEY commands in the ON KEY structures cannot be accessed while
SharkBase waits for a keystroke in the MENU( function or the WAIT command.
The following is a trivial but complete program using an ON KEY
structures to edit a file while READ is being executed on a TEXT screen:
USE test
WINDOW
CLS
dum1='1234'
dum2='ABCD'
ON key
KEY F10 EDIT
SAVE gets to test
SAVE screen to test
EDIT
dum1='2468'
RESTORE screen from test
RESTORE gets from test
ENDON
TEXT
@dum1
@dum2
TEST SCREEN LINE ....................0
TEST SCREEN LINE ....................1
ENDTEXT
DO while t
READ
ENDDO
Here is a more complex example in which an ON KEY structure is used to
add members to a data file while:
** DEMO3.PRG
SET talk off ;suppress messages
SET function off ;we want the raw function keys - not the pre-loaded
messages.
:company="United Widgets, Inc."
USE#1 order index order
USE#2 members index cust_no
SET relation on cust:no to 2
ON escape
PERFORM shutdown ;turns on function keys and cancels
ENDON
WINDOW 0,23
ON KEY
KEY F5 'NEW CUST'
SAVE gets TO temp
SAVE screen TO temp
WINDOW 0,23
SELECT 2 ;work on second file
APPEND blank ;add new member record
EDIT ;fill it in
SELECT 1 ;back to main file
REPLACE CUST:NO WITH CUST_NO#2;transfer new member number to current order
RESTORE screen temp
RESTORE gets temp
ENDON
IF :color<>7 ;test for color monitor
SET color to 31 ;white on blue
ENDIF
COLOR :color,0,0,23,79,177 ;fill screen with pattern
* 177 is a shaded fill character.
DO WHILE t ;put main menu in an infinite loop
ERASE
WINDOW 6,18,19,62 double ;declare space for menu text
mode='?'
ERASE ;fills window with blanks
SELECT 1
TEXT
Edit/Browse Demonstation
0. Exit SharkBase And return to DOS.
1. Edit Last Order.
2. Edit New Order.
3. Exit program - stay in Shark.
ENDTEXT
CURSOR 12,26 ; positions menu cursor over 1st character of 1st choice
* Note: KEY commands in the ON KEY structures cannot be accessed while
* SharkBase waits for a keystroke in the MENU( function or the WAIT command
selection=menu(3,36); six choices menu bar width 36
DO CASE
CASE selection=0
QUIT
CASE selection=1
GOTO bottom
PERFORM order_edit
CASE selection=2
APPEND blank
PERFORM order_edit
CASE selection=3
PERFORM shutdown
ENDCASE
ENDDO
*
PROCEDURE shutdown
WINDOW
ON KEY
ENDON
SET function on
CANCEL
ENDPROCEDURE shutdown
*
PROCEDURE order_edit
WINDOW 0,23
ERASE
WINDOW 6,12,19,68
TEXT orders
ON field
FIELD cust:no
FIND#2 &cust:no#1
IF lname#2=' ' ;member number not found in MEMBERS file
SAVE gets to ord ;saves get table to file ORD.GET
SCREEN 1,2 ;save current screen to internal screen buffer
WINDOW 0,23
SELECT 2 ;work on members file
NEAR#2 &cust:no#1 ;position as close as possible to right member
BROWSE ;give user chance to select correct member
SELECT 1 ;back to main file
REPLACE cust:no#1 with cust_no#2 ;transfer selected member number
SCREEN 2,1 ;restore EDIT screen
RESTORE gets from ord
ENDIF
ENDON
EDIT OFF
WINDOW 0,23
ENDPROCEDURE order_edit
*
* *** end of DEMO3.PRG ***
Introduce the optional clause in a DO CASE program structure.
OTHERWISE
In a DO CASE program structure, if no condition is true, Shark executes
the program segment following the OTHERWISE command, if any. Only one
OTHERWISE command is permitted within any DO CASE structure.
Example:
DO CASE
CASE CHOICE = 1
(some code to do something)
CASE CHOICE = 2
(some code)
CASE CHOICE = 3
(some code)
OTHERWISE
CANCEL
ENDCASE
Copy the fields from a record in one data file directly into the current record
of another data file.
OVERWRITE TO <filenum>
<filenum> any number of the data file containing the record to
be overwritten
The OVERWRITE TO command is primarily used to implement transaction
processing and to avoid lengthy record locks while records are being edited in
a network environment. Its companion command is APPEND TO.
The technique in both cases is essentially the same, utilizing a main,
indexed file and a small, unindexed personal file for each operator.
The operator edits only in the personal file, adding records as required
and verifying that everything is correct before moving the data back into the
main file. If the records already exist in the main file, OVERWRITE TO is
used to put the corrected records back where they came from; if new records
were appended to the personal file, APPEND TO is used to add them to the
end of the main file.
The two files do not need to have the same structures. Any fields which
do not exist in both files will be unaffected; any that are smaller in one
than the other create risks of losing some data due to string truncation or
numeric overflow.
Executing OVERWRITE TO is much faster than a long series of REPLACE
commands.
Example in a program with SET NETWORK ON (Shark Network Edition only):
USE customer INDEX custnum,custname SHARE
USE#2 personal
SET FUNCTION OFF
DO WHILE t
SET SAVE OFF
EDIT TEXT customer ;use formatted EDIT
IF usedby<>0
WINDOW 20,30,22,75 DOUBLE ;pop up a window for message
recnum=RECNO(1)
TEXT
.. recnum,'999,999'
.. usedby,999
Record &recnum is in use by user &usedby
Press any key to skip to next record...
ENDTEXT
junk=INKEY() ;wait to keystroke
SKIP
LOOP
ENDIF
IF :key=324 ;F10 means edit this record
REPLACE usedby with :USER ;mark record as in use by you
OVERWRITE TO 2 ;then move into personal file
SELECT 2 ;select personal file
SET SAVE ON ;allow changes to be saved
CLS ;clear the screen
TEXT custedit ;with "CHANGES PERMITTED" message
READ
WINDOW 20,30,22,75 DOUBLE ;pop up a window for confirmation
ans='No '
recnum=RECNO(1)
TEXT
.. recnum,'999,999'
..,ans,!xx
Changes may have been made to record #recnum
Save all changes (Yes/No) @ans
ENDTEXT
IF ans='Y'
OVERWRITE to 1
ENDIF
REPLACE usedby with 0 ;free the record for other users
ENDIF
ENDDO
This command removes records that have been marked for deletion by the
DELETE command or the Ctrl-U of the EDIT and BROWSE commands.
All the index files in use are automatically reindexed after PACK. (See
INDEX and REINDEX.)
Make sure you have a back up copy of the file before packing it. If PACK
is interrupted, the file may be corrupt or contain duplicate records.
PACK does not release disk space unless the TRUNCATE option is used. If a
file frequently gets large, has many records deleted and then grows again, it
may be more efficient to PACK without TRUNCATE, so that DOS is not constantly
adding disk clusters and recovering them again.
On a network with SET NETWORK ON in Shark Network Edition,
attempting to PACK or create an index while another user is accessing the data
file or an index file with the same name as any of the current indexes will
cause a LOCK error. Always LOCK the data file before attempting to PACK.
Examples:
1>USE employee
1>GO 2
1>DELETE
1 DELETE(S)
1>LIST name
1 Marek
2 *Balzer
3 Steiner
4 Rayme
5 Poyner
6 Wilson
1>PACK
5 TOTAL PACKED
1>LIST name
1 Marek
2 Steiner
3 Rayme
4 Poyner
5 Wilson
PERFORM <procedure>
<procedure> the name of the procedure
The command PERFORM is used to execute a procedure, <procedure>, defined
in the current program; <procedure> cannot be a macro. The procedure is
defined with the PROCEDURE command (see PROCEDURE).
The procedure is compiled where it is first called with PERFORM. If the
procedure makes references to fields, the structure and order of the files in
use must be the same any time the procedure is used. The same comment also
applies to memory variables. Normally, procedures perform tasks that are
independent of the environment (files in use and memory variables) or they are
always invoked in the same environment.
If you need the same procedure in different environments, write the
procedure as a separate program, and invoke it as a subroutine with the DO
command. See the command DO.
Performing procedures is a very fast way of executing the same program
segment from many places in the program. The procedures themselves must be at
the end of the program.
Example:
PERFORM header
DO WHILE .NOT. EOF
...
line=line+1
IF line>55
PERF header
ENDIF
ENDDO
CHAIN menu
*
PROCEDURE header
TEXT
&:DATE &:COMPANY Page &page
Sales Journal
Date Invoice # Description Units Dollars
ENDTEXT
line=6
ENDPROCEDURE
Post the FROM (transaction) file to the POSTING file (selected file with
index).
POST ON <str exp> FROM <file> [FIELDS ] [<scope>]
[FOR <cond>
<str exp> the expression that is the key for finding the records
in the FROM file
<file> the FROM file
Options:
FIELDS <mixed field list> the fields to be totalled in the
FROM file and added to the fields
in the POSTING file (the selected file);
<field> WITH <num exp> can add an expression
<scope> selection restricted by scope
(default scope: ALL)
FOR <cond> selection restricted by <cond>
The selected file, called the POSTING file, must be indexed by <str exp>,
an expression formed from fields of the FROM file, called the FROM file, and
constants and memory variables.
This procedure takes all the appropriate records (that is, the first
record satisfying the <cond> and within the scope) of the FROM file, one at a
time. It picks the first appropriate record, and evaluates <str exp> in the
FROM file. Next it finds the matching record in the POSTING file. If the
find is not successful, an error message POST ERROR" is sent.
If the find is successful, the procedure posts. Posting consists of one
of two actions, depending on whether a field name is listed alone in the mixed
field list, or if a field name has a WITH clause attached.
1. If the field is named alone, the value of that field in the FROM file is
added to the value already in the POSTING file.; that is, it adds the fields
in the <field list> of the current record of the FROM file to the fields of
the same name in the current record of the POSTING file.
2. In a mixed field list, the clause <field> WITH <num exp> allows adding any
numeric expression <field> to the field in the POSTING file. <num exp> is
formed using the fields of the FROM file, and optionally fields from other
files (including the POSTING file), variables, and constants.
If a find is unsuccessful and SET ADD ON, a record will be appended
to the POSTING file, the non-numeric fields of the POSTING file record will
be replaced by the matching fields in the FROM file record, and, finally,
the procedure posts as above.
The procedure continues with the next appropriate record of the FROM file.
If there is no FIELDS clause, POST simply checks whether there is a
record in the POSTING file for each value of <str exp> in the FROM file. This
is valuable to check whether every record in the FROM file has a matching
record in the POSTING file without changing any fields.
If SET PRINT OFF (the default), the error messages appear on the screen;
otherwise, they are echoed to (printed on) the printer (or to the alternate
text file, if SET ALTERNATE ON).
Deleted records are not posted, irrespective of SET DELETED.
Important programming note: The POST command automatically opens the FROM
file in its internal work area; if the FROM file is already open in another
work area, the compiler will assume it is closed when the POST command is
passed during execution, even if that command is in an IF, CASE or other
structure module that is not executed!
Therefore, if there is any reference to the FROM file later in the
program, open the file in the proper work area again immediately after the POST
command. If it is not actually needed after the POST is executed (e.g., the
program exits after the POST), open it with the COMPILE keyword. Example:
USE#4 invoices COMPILE
See also the related commands: TOTAL and UPDATE.
Example:
Use the data files SALESMEN and ORDER (see Appendix), to record the amount
of merchandise sold.
1>SET TALK ON
1>USE salesmen
1>INDEX ON slsman:no TO salesmen
2 RECORDS IN TOTAL INDEXED
1>POST ON slsman:no FROM order FIELDS amount FROM order
2 ** POST ERROR ** 4
5 POST(S)
1>LIST
1 1 Smith John 4386.00
2 5 Willson Tom 14679.50
The error message POST ERROR indicates that in the ORDER file there is a
record (number 2) with a salesman code ("4") not found in the SALESMEN file.
Note that the amount for Smith is 2709.00 (the amount that was in the
SALESMEN file) +1677.00 (the amount in ORDER record 5).
The amount for Willson is 12089.00 (the amount that was in the SALESMEN
file) +597.50 (the amount in order 1) +559.00 (the amount in order 3) +1434.00
(the amount in order 4).
Of course, the most typical example of the use of the command POST is
posting to a general ledger (the POSTING file) from a journal (the transaction
file).
Procedures should be used for all Shark program segments that
are used more than once in a program. However, if the procedure makes
references to fields, the structure and order of the files in use must be the
same any time the procedure is used. The same comment also applies to memory
variables. Normally, procedures perform tasks that are independent of the
environment (files in use and memory variables) or they are always invoked in
the same environment.
If the whole program will not fit in memory, then replace the procedures
by subroutines invoked by the DO command. (See CHAIN, PERFORM, and PROCEDURE.)
The procedure name, <procedure>, follows the same rules as variable names
(see Section 2); <procedure> cannot be a macro. Procedures must be placed at
the end of the program. A procedure can perform other procedures.
A procedure is compiled at the first PERFORM invoking it. (See Example 1
in Appendix A)
Hint: If the same procedure must be used in more than one subroutine, add
a digit (1, 2, 3, etc.) to the name in the various subroutines. Example: if a
procedure called BELL is needed in main program ENTER and subroutines ENTER1
and ENTER2, call it BELL, BELL1 and BELL2 in the three program modules
respectively.
Examples:
1. A simple label printing procedure.
INPUT 'Enter the number of labels to print ' TO nlabel
REPEAT nlabel TIMES
PERFORM label
SKIP
ENDREPEAT
*
PROCEDURE label ; prints 5 lines of data and 4 blank lines
? name ; on standard labels one across
? add:1
? add:2
? add:3
? postal
?
?
?
?
ENDPROCEDURE ; end of label printing
2. A procedure to clear screen, make border, set standard heading, and
create reverse video black for menu choices. This procedure would be called
before any menu display in the program.
PROCEDURE MENUSCREEN
ERASE
WINDOW 2,1,23,77 DOUBLE
COLOR 112, 6, 4, 22, 74
@ 3,0 SAY CEN(:COMPANY, 80)
@ 5,0 SAY CEN('Enter your choice below:',80)
ENDPROCEDURE menuscreen
The PSTRING command sends a printer-control string to the printer without itself printing. It is used for styling printer fonts in Shark documents.
PSTRING = a list of characters, in a special format, to be output to the printer or the printer file
An enormous variety of printers emerged under MS-DOS, with their often-obtuse
control languages such as HP's PCL (printer control language). Programming languages like Shark needed to pass these instructions embedded in text output in order to control attached printers.
SharkBase offers a limited set of tools to specify characters in your printouts.
The setup string is a list of character descriptors, separated by
commas. Each descriptor may be a number (the ASCII value of a character), a hex number followed by an "h" (the hex value of a character), or any alpha character.
Examples:
PSTR 27,W,1Bh,41h,3h
A variable or field containing the setup string may be used in a macro:
ps='27,W,1Bh,41h,3h'
PSTRING &ps
These macros can be stored in a text file which is read by a Shark program along with a text document that is to be printed.
In order to be fed properly to the printer, the macros need to be parsed by a Shark program and stored in memory at the time of printing. They are read into memory with a small program such as this:
* read in printer data macros from PRINTER.DTA ...
OK=ROPEN("PRINTER.DTA",2); opens the printer PCL data file
IF .NOT. OK
RING
? "Printer not installed!"
WAIT
CANCEL
ELSE
OK=READ(_PNORM,2)<-- normal text
OK=READ(_PCOMP,2)<-- compressed
OK=READ(_PCOFF,2)<-- compressed off
OK=READ(_PBOLD,2)<-- bold
OK=READ(_PBOFF,2)<-- bold off
OK=READ(_PWIDE,2)<-- wide
OK=READ(_PWOFF,2)<-- wide off
OK=READ(_PRSET,2)<-- reset printer
OK=READ(_PFF,2)<-- form feed (not used in vDOS/Windows)
ENDIF
*release printer
OK=CLOSE(2)
vDOS prints Shark text output without PCL codes, but the printout will be limited to your single basic default font, likely the Windows "Courier" font. PSTRING allows some limited management of your default font, mainly character weight (light, bold, etc).
QUIT
Option:
a value passed to DOS, which allows batch files
to control their execution with the
IF ERRORLEVEL command
The command QUIT exits from Shark. It closes all the files and
gives you the prompt of the operating system.
If Shark is running unattended in a batch file, a program can
pass an "errorlevel" to DOS indicating if some condition was encountered
during execution. For instance, if there is an error reading a data file,
you can QUIT to DOS with an errorlevel of 1, which the batch file can detect
so that the tape backup is not run. The command in this case wound be QUIT 1.
Consult your DOS reference manual for the IF ERRORLEVEL batch program command.
Example:
DO CASE
CASE ans='0'
QUIT
CASE ans='2'
CHAIN prog1
...
OTHERWISE
QUIT 1 ;tells DOS an error encountered by Shark
ENDCASE
Enter full-screen editing and data entry specified by the @ GET and TEXT
commands.
READ
The READ command activates the input aspect of the @ GET commands
(including the @ and % macros in a TEXT). The @ GET command displays the
variable at the specified location and format. The READ command puts the
screen in full-screen editing mode with the current values of all fields and
memory variables shown, and the new values of all the variables in the @ GET
(and the @ and % macros in TEXT) commands can be entered.
If the READ changes any key of a record in an index file in use, then the
index files are updated.
The READ command clears all the pending GET commands.
For the use of the editing keys in full-screen editing.
Example:
@ SAY 20,0 'Enter your choice: ' GET ans
READ
is the same as:
ACCEPT 'Enter your choice: ' TO ans
except that @ SAY does not create ANS, ACCEPT does; @ SAY directs the query to
a specific location on the screen.
Of course, the real use of the @ SAY and READ commands is to display more
than one field for input.
Undelete selected deleted records in the selected file.
RECALL <scope> [FOR <cond>]
Options:
<scope> select by scope (default scope: current record)
FOR <cond> select by condition
Records are flagged as DELETED either with Ctrl-U in BROWSE, or EDIT, or
with the DELETE command (see the commands BROWSE, DELETE, and EDIT). The
command RECALL reverses the DELETED flag.
The command PACK physically removes all records flagged for deletion.
After a PACK, the records can no longer be recalled.
Note that if SET DELETE ON, then RECALL will work only on a deleted
record only if it is also the current record. Normally SET DELETE OFF
before using RECALL.
Examples:
1>USE employee
1>GO 2
1>DELETE
1 DELETE(S)
1>LIST name
1 Marek
2 *Steiner
3 Rayme
4 Poyner
5 Poyner
6 Wilson
1>GO 2
1>RECALL
1 RECALL(S) ;record 2 was recalled
1>RECALL ;the default is the current record
0 RECALL(S)
1>DELETE RECORD 2
1 DELETE(S)
1>DELETE RECORD 4
1 DELETE(S)
1>DELETE RECORD 6
1 DELETE(S)
1>RECALL ALL
3 RECALL(S) Note: Records 2, 4, 6 are recalled
1>GO TOP
1>DELETE
1 DELETE(S)
1>DELETE RECORD 3
1 DELETE(S)
1>DELETE RECORD 5
1 DELETE(S)
1>LIST name
1 *Marek
2 Steiner
3 *Rayme
4 Poyner
5 *Poyner
6 Wilson
1>GO TOP
1>RECALL NEXT 4
2 RECALL(S) Note: Records 1 and 3 are recalled
This command takes all the index files currently in use for the selected
data file, and rebuilds each index file.
In Conversational Shark, and in programs with SET TALK ON,
REINDEX will output the key, the progress of the indexing procedure, showing
the completion of each 100 records, and the total number of records indexed.
A FOR clause can be used to limit the records in the index(es) to those
for which <cond> is true. See INDEX.
On a network with SET NETWORK ON in Shark Network Edition,
attempting to REINDEX while another user is accessing an index file with the
same name as one of the current file's indexes will cause a LOCK error.
Examples:
The commands:
1>USE employee
1>INDEX on name+fname TO employee
accomplish the same as the following:
1>USE employee INDEX employee
1>REINDEX
Note that for REINDEX, you do not have to specify the key (name+fname) or
the name of the index file (EMPLOYEE.NDX).
In a program:
USE employee INDEX employee,empl1,empl2
REINDEX
will rebuild three index files: EMPLOYEE.NDX,EMPL1.NDX,EMPL2.NDX.
The following uses a FOR clause to limit the index to current employee
only, based on date of termination being blank:
USE employee INDEX employee,empl1,empl2
REINDEX FOR terminated=' '
<memvar list> the memory variables to be released,
separated by commas
EXCEPT <memvar list> the memory variables not to be released, separated by commas
ALL release ALL
The command RELEASE erases memory variables; it releases memory space for
new memory variables. You can make a list of the memory variables you want
removed, use the parameter ALL to remove them all, or use the parameter EXCEPT
<memvar list> to remove all except those on the list. RELEASE ALL may
cripple your application if your application loops back to the beginning (the
variables required will all be gone).
This command should be used in programs only with the utmost care.
Compiled programs set pointers to memory variables. If the memory variables
are released, the release must be absolute, not dependent on some conditions.
Since 128 memory variables can be used at any one time, and any number of
them may be matrix variables, the use of RELEASE during a program may require careful
debugging afterward! When programs are chained, Shark will do a CLEAR, which
includes a RELEASE ALL.
RELEASE does not affect the values of the function keys and the system variables.
Examples:
1. In conversational Shark:
1>LIST MEMO
Name Type Width Contents
** Total ** 0 variables, 0 bytes
1>name='David'
1>age='11'
1>LIST MEMO
Name Type Width Contents
NAME C 5 David
AGE C 2 11
** Total ** 2 variables, 7 bytes
1>RELEASE age
1>LIST MEMO
Name Type Width Contents
NAME C 5 David
** Total ** 1 variables, 5 bytes
1>RELEASE ;RELEASE by itself does nothing
1>LIST MEMO
Name Type Width Contents
NAME C 5 David
** Total ** 1 variables, 5 bytes
1>age=11
1>LIST MEMO
Name Type Width Contents
NAME C 5 David
AGE N 8 11
** Total ** 2 variables, 13 bytes
1>RELEASE ALL
1>LIST MEMO
Name Type Width Contents
** Total ** 0 variables, 0 bytes
2. In programming, from a program you might DO a number of subroutines;
each uses a number of memory variables, most of which are unique to it, but
which in total exceed the 128 memory variable limit. In this case, decide
which variables needed to be retained, and use the RELEASE EXCEPT after each
subroutine as follows:
@ 20,30 SAY ' Enter your choice: ' GET choice PICTURE'!'
DO CASE choice ;menu choices
CASE choice='A'
DO proga
RELEASE EXCEPT choice,cust,due,amount
CASE choice='B'
DO progb
RELEASE EXCEPT choice,cust,due,amount
CASE choice='C'
DO progc
RELEASE EXCEPT choice,cust,due,amount
ENDCASE
ENDDO ;menu choices
RENAME <file> TO <new file>
<file> the file name to be changed (default extension DBF)
<new file> the new name for the file (default extension DBF)
This command is similar to the RENAME command of the operating system: a
file is renamed. No wild card characters are allowed.
If the file name has no extension, the extension DBF is assumed. (Do not
rename a data or index file to a name starting with the letters COMP;
Shark is unable to open such a file.)
Examples:
1>RENAME employee TO empl
Renames EMPLOYEE.DBF to EMPL.DBF
1>RENAME employee.dbf TO employee.xyz
RENAME FIELD <field> TO <new fieldname>
<field> the field name to be changed
<new fieldname> the new name for the field
This command allows you to change the name of a field in a data file
without doing a MODIFY, which potentially endangers the file and removes all
deleted records, often necessitating a REINDEX.
Examples:
1>RENAME FIELD employee TO empl
Renames the field EMPLOYEE to EMPL in the current data file.
1>RENAME#3 FIELD cust TO customer
Caution: All programs, report forms and index files that reference the old
field name must be revised before they will work properly after a field is
renamed.
Repeat program segment a specified number of times.
REPEAT <num exp> TIMES [VARYING <num var>]
<num exp> a numeric expression; repeat the loop this many times
Option:
VARYING <num var> loop counter
The REPEAT command introduces the program structure:
REPEAT <num exp> TIMES [VARYING <num var>]
<program segment>
ENDREPEAT
The program lines in the <program segment> are repeated <num exp> times.
The end of the program segment is indicated by ENDREPEAT.
The optional VARYING clause is used to vary a memory variable from 1 to
<num exp>; at the start of the first loop, <num var> is initialized as 1;
<num var> is incremented by 1 each time the top of the loop is reached.
This command is especially useful to manipulate matrix variables.
Repeats can be nested.
The BREAK command is used in a REPEAT loop to exit at the bottom of the
loop. If there are nested loops, the exit is at the bottom of the innermost
loop then being executed.
The Shark keywords: TIMES and VARYING should not be in macros.
<num var> cannot be a matrix variable.
Note that, if <program segment> contains commands that move the current
record pointer (for instance, a SKIP command), REPEAT does not terminate if
EOF becomes true. Thus displaying a record and skipping in a file with 50
records, using REPEAT 100 TIMES, will display the last record in the file
51 times. DO WHILE .NOT. EOF is the better construction for such a task.
When editing with the internal Shark programming editor (see
WRITE command), Alt-F reformats the file with all structures properly indented,
making it easy to see unbalanced structures.
Examples:
SET WIDTH TO 40
DIM NUM num[10]
REPEAT 10 TIMES VARYING count
num[count]=POW(2,count)
ENDREPEAT
? num
REPLACE <scope> <field1> WITH <exp1> [,<field2> WITH <exp2> ... ]
[FOR <cond>]
Options:
<scope> select by scope
(default scope: current record)
<field> WITH <exp>... list fields and expressions for replacement
FOR <cond> select by condition
This command is used for changing some records in the selected data file.
The number of <field> WITH <exp> is only limited by the length of a command
line (254 characters).
The scope ALL can be used to change all the records in the data file.
This could be dangerous. Consider backing up your data file before you use
this command.
This command is very useful, especially in conjunction with SET RELATION
TO, so that the fields from which <exp> is formed may come from a number of
data files.
Examples:
1>USE employee
1>REPLACE name WITH 'Mareck'
1>? name
Mareck
1>REPLACE ALL salary WITH salary*1.05
6 REPLACE(S)
1>LIST name,salary
1 Mareck 27090.00
2 Steiner 37569.00
3 Rayme 83065.50
4 Poyner 9145.50
5 Poyner 9145.50
6 Wilson 35563.50
1>GO TOP
1>REPLACE NEXT 4 salary WITH salary*1.06 FOR name < 'P'
1 REPLACE(S)
1>USE order
1>USE#2 clothing
1>SET RELATION ON inven:no TO 2
1>REPLACE price WITH price#2
To replace more than one field:
REPLACE q1 WITH q1-q1#2,q2 WITH q2-q2#2,q3 WITH q3-q3#2
