In WORKING-STORAGE you will need a field that will be used to indicate when the end of file is reached. Trying to read beyond the end of a file is disastrous. Define a field similar to:
01 WS-END-OF-FILE-SW PIC X VALUE 'N'.
88 WS-END-OF-FILE VALUE 'Y'.
A skeleton of the PROCEDURE DIVISION will be similar to this:
PROCEDURE DIVISION.
PERFORM 100-INITIALIZE.
PERFORM 200-PROCESS-INPUT
UNTIL WS-END-OF-FILE.
PERFORM 900-TERMINATE.
STOP RUN.
100-INITIALIZE.
* THE FOLLOWING STATEMENT CAN BE USED TO OPEN ALL FILES
OPEN INPUT INPUT-FILE.
* DO OTHER INITIALIZATION TASKS IN THIS PARAGRAPH, LIKE
* GETTING CURRENT DATE, INITIALIZING VARIABLES, ETC.
* THE PERFORM OF THE READ PARAGRAPH SHOULD BE THE LAST STATEMENT
* IN THIS PARAGRAPH. THIS IS THE PRIMING READ.
PERFORM 110-READ-INPUT-FILE.
110-READ-INPUT-FILE.
* IF MULTIPLE INPUT FILES HAVE A SEPARATE READ PARAGRAPH
* FOR EACH. ALSO HAVE SEPARATE END-OF-FILE INDICATORS.
READ INPUT-FILE
AT END MOVE 'Y' TO WS-END-OF-FILE-SW
END-READ.
200-PROCESS-INPUT.
* ENTER CODE HERE TO PROCESS THE INPUT DATA
* THE PERFORM OF THE READ PARAGRAPH SHOULD BE THE LAST STATEMENT
* IN THIS PARAGRAPH.
PERFORM 110-READ-INPUT-FILE
900-TERMINATE.
* THE FOLLOWING STATEMENT CAN BE USED TO CLOSE ALL FILES
CLOSE INPUT-FILE.
File matching is also known as co-sequential file processing, because multiple input files are being processed at once. All files must be sorted on the field(s) to be matched.
For example we have two input files, IN-ORDER and IN-CUSTOMER and we need to match them by customer number, a field appearing on both files. The basic logic will be similar to this:
PERFORM 200-READ-ORDER.
PERFORM 210-READ-CUSTOMER.
PERFORM 300-COMPARE
UNTIL END-OF-CUSTOMER AND
END-OF-ORDER.
STOP RUN.
200-READ-ORDER.
READ IN-ORDER
AT END
MOVE 'Y' TO END-OF-ORDER-SW
MOVE HIGH-VALUES TO O-CUST-NBR
END-READ.
210-READ-CUSTOMER.
READ IN-ORDER
AT END
MOVE 'Y' TO END-OF-ORDER-SW
MOVE HIGH-VALUES TO C-CUST-NBR
END-READ.
300-COMPARE.
IF O-CUST-NBR = C-CUST-NBR
PERFORM P400-MATCH
PERFORM P200-READ-ORDER
PERFORM P210-READ-CUSTOMER
ELSE
IF O-CUST-NBR < C-CUST-NBR
PERFORM P500-ORDER-W-NO-CUSTOMER
PERFORM P200-READ-ORDER
ELSE
PERFORM P600-CUSTOMER-W-NO-ORDER
PERFORM P210-READ-CUSTOMER
END-IF
END-IF.
Notes:
Sorting a file requires defining two files in addition to the input (unsorted) file: the sorted file and a work file. Each requires a SELECT statement. The sorted file requires an FD and the work file requires an SD. SDs (Sort file Description) are just like FDs but cannot have a LABELS clause.
SELECT IN-FILE ASSIGN TO DISK.
SELECT SORTED-FILE ASSIGN TO DISK.
SELECT SORT-WORK ASSIGN TO DISK.
FD IN-FILE
VALUE OF FILE-ID IS 'PAYROLL-FILE'.
01 IN-RECORD PIC X(200).
FD SORTED-FILE
VALUE OF FILE-ID IS 'SORTED-PAYROL-FILE'.
01 SORTED-RECORD PIC X(200).
* ONLY NEED TO DEFINE THE FIELDS THAT THE FILE WILL BE
* SORTED ON. BUT ACCOUNT FOR THE ENTIRE RECORD.
SD SORT-WORK
VALUE OF FILE-ID IS 'TEMP-1'.
01 SORT-RECORD.
05 SR-EMPLOYEE PIC 9(09).
05 PIC X(21).
05 SR-REGION PIC X(03).
05 PIC X(87).
05 SR-DEPT PIC 9(02).
05 PIC X(78).
The above SD has prepared us to sort on the employee number, region and department number. Let's sort on region first, then department and then employee.
One COBOL statement will do the trick. The SORT statement will read the entire input file and sort it. You will use the sorted file as input to the rest of the program. Treat it as any other input file (read it, check for end of file, close it, etc.). Normally, the sort is done during program initialization but this depends on processing needs.
SORT SORT-WORK
ON ASCENDING KEY SR-REGION
ON ASCENDING KEY SR-DEPT
ON ASCENDING KEY SR-EMPLOYEE
USING IN-FILE
GIVING SORTED-FILE.
* SOME COMPILERS CLOSE THE FILE AFTER THE SORT, SOME DO NOT.
* YOU MAY OR MAY NOT HAVE TO OPEN IT BEFORE CONTINUING.
OPEN INPUT SORTED-FILE.
PERFORM UNTIL WS-EOF-SW = 'Y'
READ SORTED-FILE
AT END
MOVE 'Y' TO WS-EOF-SW
NOT AT END
PERFORM 200-PROCESS
END-READ
END-PERFORM.
CLOSE SORTED-FILE.
STOP RUN.
As an alternative to specifying input and output files in the SORT you can instead specify an INPUT PROCEDURE or OUTPUT PROCEDURE, paragraphs to execute before and after the sort takes place.
SORT SORT-WORK
ON ASCENDING KEY SR-REGION
ON ASCENDING KEY SR-DEPT
ON ASCENDING KEY SR-EMPLOYEE
INPUT PROCEDURE 100-PRE-SORT
OUTPUT PROCEDURE 200-POST-SORT.
100-PRE-SORT.
OPEN INPUT IN-FILE.
PERFORM UNTIL WS-EOF-SW = 'Y'
READ IN-FILE
AT END
MOVE 'Y' TO WS-EOF-SW
NOT AT END
PERFORM 110-PROCESS
END-READ
END-PERFORM.
CLOSE IN-FILE.
110-PROCESS.
MOVE IN-RECORD TO SORT-RECORD.
* ONLY SORT RECORDS WITH NON-ZERO DEPARTMENT NUMBER
IF SR-DEPT NOT EQUAL ZEROES
RELEASE SORT-RECORD
END-IF.
200-POST-SORT.
OPEN INPUT SORT-FILE.
MOVE 'N' TO WS-EOF-SW.
PERFORM UNTIL WS-EOF-SW = 'Y'
RETURN SORT-FILE
AT END
MOVE 'Y' TO WS-EOF-SW
NOT AT END
PERFORM 300-PROCESS
END-READ
END-PERFORM.
CLOSE SORT-FILE.
In WORKING-STORAGE define something similar to this (the lines per page will depend on the installation. Initialize the line number to the same value as lines per page to trigger the printing of headings for the first page):
01 WS-REPORT-VALUES.
05 WS-PAGE-NBR PIC 9(4).
05 WS-LINE-NBR PIC 99 VALUE 66.
05 WS-LINES-PER-PAGE PIC 99 VALUE 66.
Also in WORKING-STORAGE define separate records for each page heading line, each column heading line, each detail line and any total lines. The record layout in the FD for the report file should not contain any fields, only the record name and a PIC X big enough to hold any of the heading or detail records plus 1 character. If there are to be totals produced for the report then an accumulator field for each total must also be defined in WORKING-STORAGE (make sure it's big enough).
Most reports contain the current date in one of the page headings. If so, get the current date in the program's initialization paragraph and store it on the page heading record. Do not get the current date every time the headings are printed. If your detail lines require special initializations before they are printed then call that paragraph during the program's initialization paragraph.
While processing an input record you will be moving data to the detail
line. At the end of the processing loop, but before reading the next input
record, call the print detail line paragraph (the detail line now has all
of the data that will be printed).
8000-PRINT-DETAIL.
* IF ANY OF THE FIELDS ON THE DETAIL LINE ARE TO BE TOTALLED THEN
* ADD ITS VALUE TO THE ACCUMULATING FIELD HERE, BEFORE ANYTHING
* IS PRINTED.
ADD 1 TO WS-LINE-NBR.
IF WS-LINE-NBR > WS-LINES-PER-PAGE
PERFORM 8100-PRINT-HEADINGS
END-IF.
WRITE OT-REPORT-RECORD
FROM WS-DETAIL-RECORD
AFTER ADVANCING 1 LINE
END-WRITE.
* IF THE DETAIL LINES ARE NOT TO BE SINGLE SPACED THEN CHANGE
* THE NUMBER OF LINES TO ADVANCE IN PREVIOUS WRITE AND ALSO THE
* VALUE IN THE FOLLOWING ADD TO REFLECT DESIRED SPACING.
ADD 1 TO WS-LINE-NBR.
* IF YOUR DETAIL LINES REQUIRE SPECIAL INITIALIZATIONS THEN CALL
* THAT PARAGRAPH AT THE END OF THIS PARAGRAPH
8100-PRINT-HEADINGS.
ADD 1 TO WS-PAGE-NBR
* MOVE ANY FIELDS THAT ARE PART OF THE PAGE HEADINGS TO THE
* PAGE HEADING RECORDS NOW (LIKE THE PAGE NUMBER).
WRITE OT-REPORT-RECORD
FROM WS-PAGE-HDR-1-RECORD
AFTER PAGE
END-WRITE.
* INCLUDE SIMILAR WRITES FOR ALL REMAINING PAGE HEADING RECORDS AND
* COLUMN HEADING RECORDS - EACH ADVANCING THE PROPER NUMBER OF LINES
* TO REFLECT DESIRED SPACING (SINGLE, DOUBLE, ETC.).
*
* THE WS-LINE-NBR FIELD WILL NEED POPULATED WITH A VALUE INDICATING
* HOW MANY LINES HAVE BEEN PRINTED ON THE PAGE DURING THIS PARAGRAPH,
* INCLUDING SPACING.
MOVE literal TO WS-LINE-NBR.
If any totals were accumulated during the generation of the report then call the paragraph to print the total line(s) from the program's termination paragraph (before you close the report file).
It is often valid for a report to have no data. Either the input file is empty (not necessarily a problem) or the input file has no records meeting whatever criteria is necessary in order to get printed on the report. The problem is that in such cases nothing comes out of the program, not even the headings (In the algorithm above the 1st headings aren't printed until the 1st detail line is built).
How does the user know that the report is empty and there wasn't some problem that crashed the program? A common practice is to print a descriptive message on the report signifying that there was no data to print so that the end user knows that this was the case. Such a message needs defined in WORKING-STORAGE:
01 WS-DETAIL-0.
05 PIC X(54) VALUE SPACES.
05 PIC X(24) VALUE '*** NO DATA TO PRINT ***'.
05 PIC X(54) VALUE SPACES.
In the program's termination paragraph check the value of the page number field. If it is still zero then you haven't printed any detail records. Call the paragraph to print the page headings, then print the message to signify that there was no data print. If printing totals on the report then have the call to that paragraph as an 'ELSE' to the 'IF WS-PAGE-NBR = 0'.
Click here to view/download txt file with control break algorthims (single and double level).
In WORKING-STORAGE define fields to hold the current values of each field that is a control field. They must be defined exactly as the corresponding input fields are defined. Initialize them to ZEROS if numeric, LOW-VALUES if alphanumeric.
Special processing is done whenever the value of a control field changes. Totals must be produced for the previous control group, fields must be initialized for the new control group and possibly some headings or a new page for the new control group. Checks for control break processing are done immediately after a record is read. Compare the value of the control field on the input record to the value in the hold field. If different then call the paragraph that handles control break processing.
In WORKING-STORAGE:
01 WS-CONTROL-FIELDS.
05 WS-HOLD-DEPARTMENT PIC X(03) VALUE LOW-VALUES.
05 WS-FIRST-BREAK-SW PIC X VALUE 'Y'.
88 WS-FIRST-BREAK VALUE 'Y'.
In the PROCEDURE DIVISION:
2000-READ-INPUT-RECORD.
READ INPUT-FILE
AT END
MOVE 'Y' TO WS-END-OF-FILE-SW
NOT AT END
PERFORM 2100-CHECK-FOR-BREAK
END-READ.
2100-CHECK-FOR-BREAK.
IF IN-DEPARTMENT NOT = WS-HOLD-DEPARTMENT
* IF THERE WAS NO PREVIOUS CONTROL GROUP THEN DON'T TRY TO TOTAL
* ANYTHING OR PRINT TOTAL LINE(S).
IF WS-FIRST-BREAK
MOVE 'N' TO WS-FIRST-BREAK-SW
ELSE
PERFORM 8100-TOTAL-PREVIOUS-GROUP
END-IF
* BOTH OF THE FOLLOWING PROCESSING MAY NOT BE NECESSARY
PERFORM 8200-INITIALIZE-FOR-NEW-GROUP
PERFORM 8300-HEADINGS-NEW-GROUP
MOVE IN-DEPARTMENT TO WS-HOLD-DEPARTMENT
END-IF.
The above code is for a single-level control break. Multiple-level control breaks take more care. The control fields will be ordered (i.e. break on salesperson within a region).
Hold fields must be defined for each control fields. It may also be easier (and cleaner) to have separate paragraphs for totalling and headings for each control field. When checking for changes in the control fields start with the major field and then work your way to the most minor field (i.e. if the breaks were 'a' within 'b' within 'c' within 'd' within 'e' check field 'e' first, then 'd', etc.). A control break in a field implies a break in all fields minor to the one with the break. Ensure all the control break processing that needs to be executed actually happens and happens in the correct order.
We have a file containing part numbers and descriptions. Each record contains a 9-digit part number and a 41-byte description. This file is to be loaded into a table to be cross-referenced with a part number on another file.
Since we want the entire table to be ready when we start processing our primary input file the table will be loaded, in its entirity, during program initialization.
FD PART-MASTER.
01 PART-MASTER-REC.
05 IN-PART-NBR PIC 9(9).
05 IN-PART-DESC PIC X(41).
(In working storage)
01 FILLER.
05 PART-TABLE OCCURS 100 TIMES.
10 PART-NBR PIC 9(9).
10 PART-DESC PIC X(41).
01 WS-EOF-SW PIC X VALUE 'N'.
88 END-OF-FILE VALUE 'Y'.
88 TBL-OVERFLOW VALUE 'F'.
01 WS-SUB PIC 999.
01 WS-TBL-MAX PIC 999 VALUE 100.
(To load table from file)
P100-LOAD-TABLE.
OPEN INPUT PART-MASTER.
MOVE 0 TO WS-SUB.
PERFORM UNTIL END-OF-FILE OR TBL-OVERFLOW
READ PART-MASTER
AT END SET END-OF-FILE TO TRUE
NOT AT END PERFORM P110-ADD-ENTRY
END-READ
END-PERFORM.
CLOSE PART-MASTER.
IF TBL-OVERFLOW
DISPLAY 'ERROR: PART TABLE OVERFLOW!!!!'
DISPLAY 'PROGRAM ENDING...'
STOP RUN
END-IF.
P110-ADD-ENTRY.
ADD 1 TO WS-SUB.
IF WS-SUB > WS-TBL-MAX
SET TBL-OVERFLOW TO TRUE
ELSE
MOVE PART-MASTER-REC TO PART-TABLE (WS-SUB)
END-IF.
Now that the table has been populated with values using the previous algorithm let's try to use this table. The primary input file to this program (the sales file) contains, among other things, a part number. This is to be printed on a sales report along with the description of the part. The description is not kept on the sales file so the table is needed.
A file matching algorithm is not used becuase that would require the sales file to be sorted by part number and it is doubtful that we would want it that way on the report.
There are many ways to do this. If the table were indexed the SEARCH command could be used. The following is not using the SEARCH statement:
PERFORM P200-READ-SALES-FILE.
MOVE '*** NO PART DESCRIPTION' TO RPT-DTL-PART-DESC.
PERFORM VARYING WS-SUB FROM 1 BY 1
UNTIL (WS-SUB > WS-TBL-MAX) OR
(IN-PRT-NBR = PART-NBR (WS-SUB))
IF IN-PRT-NBR = PART-NBR (WS-SUB)
MOVE PART-DESC TO RPT-DTL-PART-DESC
END-IF
END-PERFORM.
(if the table was defined as INDEXED BY WS-INDX)
PERFORM P200-READ-SALES-FILE.
SET WS-INDX TO 1.
SEARCH PART-TABLE
AT END
MOVE '*** NO PART DESCRIPTION' TO RPT-DTL-PART-DESC
WHEN IN-PRT-NBR = PART-NBR (WS-INDX)
MOVE PART-DESC TO RPT-DTL-PART-DESC
END-SEARCH.
The SELECT for relative files requires the following clauses:
ORGANIZATION IS RELATIVE
ACCESS IS DYNAMIC
RELATIVE KEY IS working-storage-field-1
FILE STATUS IS working-storage-field-2.
working-storage-field-1 must be numeric, working-storage-field-2 is a PIC X(2).
Before a read or write the relative key must be populated with a value by the program. This is basically a record number. When a read is executed the value of the relative key directs the system to attempt a retrieval of that particular record (if relative key is 5 then the system will try to retrieve a record from the 5th slot in the file). If a write is executed then the system will attempt to place a record in that slot.
The value of the file status after the statement is executed will indicate whether or not the statement successfully executed. If not, the value will also indicate what went wrong. This field can be interrogated with an IF or EVALUATE:
SELECT IN-MASTER
ORGANIZATION IS RELATIVE
ACCESS IS DYNAMIC
RELATIVE KEY IS WS-RELATIVE-KEY
FILE STATUS IS WS-FILE-STATUS.
. . . . . .
MOVE WS-REL-ADDR TO WS-RELATIVE-KEY.
READ IN-MASTER.
EVALUATE WS-FILE-STATUS
WHEN '00'
PERFORM 300-SUCCESSFUL-READ
WHEN '23'
PERFORM 400-RECORD-SLOT-IS-EMPTY
WHEN '92'
DISPLAY 'FILE NOT OPEN'
WHEN OTHER
DISPLAY 'UNEXPECTED FILE ERROR: ' WS-FILE-STATUS
END-EVALUATE.
See the File Errors Page for a list of file status values and their meanings.
As an alternative the INVALID KEY clause can be used:
MOVE WS-REL-ADDR TO WS-RELATIVE-KEY.
READ IN-MASTER
INVALID KEY
PERFORM 400-RECORD-SLOT-IS-EMPTY
NOT INVALID KEY
PERFORM 300-SUCCESSFUL-READ
END-READ.
This lumps all unsuccessful operations into one group so this may not be an appropriate method.
Writes work similarly:
SELECT OT-MASTER
ORGANIZATION IS RELATIVE
ACCESS IS DYNAMIC
RELATIVE KEY IS WS-RELATIVE-KEY
FILE STATUS IS WS-FILE-STATUS.
. . . . . .
MOVE WS-REL-ADDR TO WS-RELATIVE-KEY.
WRITE OT-MASTER-RECORD.
EVALUATE WS-FILE-STATUS
WHEN '00'
ADD +1 TO WS-NBR-RECORDS-WRITTEN
WHEN '22'
PERFORM 500-SLOT-ALREADY-USED
WHEN '92'
DISPLAY 'FILE NOT OPEN'
WHEN OTHER
DISPLAY 'UNEXPECTED FILE ERROR: ' WS-FILE-STATUS
END-EVALUATE.
Writes also support INVALID KEY/NOT INVALID KEY.
Relative files can be accessed as if they were sequential files. In the SELECT change the ACCESS IS DYNAMIC to ACCESS IS SEQUENTIAL. Can then use READs (include AT END) and WRITEs as if the file was sequential. Do not specify a RELATIVE KEY or a FILE STATUS.
Relative files support two other type of operations, but these require the file to be opened for input and output at the same time. Access cannot be sequential. These operations are DELETE and REWRITE.
The DELETE statement will delete the current record. This requires a successful READ before the DELETE.
OPEN I-O IN-MASTER.
. . . . . .
MOVE WS-REL-ADDR TO WS-RELATIVE-KEY.
READ IN-MASTER.
EVALUATE WS-FILE-STATUS
. . .
END-EVALUATE.
DELETE IN-MASTER RECORD.
EVALUATE WS-FILE-STATUS
WHEN '00'
PERFORM 200-SUCCESSFUL-DELETE
WHEN '43'
DISPLAY ' NO CURRENT RECORD. REL KEY = ' WS-RELATIVE-KEY
. . .
END-EVALUATE.
The REWRITE statement will update the current record. This requires a successful READ before the REWRITE.
OPEN I-O IN-MASTER.
. . . . . .
MOVE WS-REL-ADDR TO WS-RELATIVE-KEY.
READ IN-MASTER.
EVALUATE WS-FILE-STATUS
. . .
END-EVALUATE.
* IF THE READ WAS SUCCESSFUL YOU WILL UPDATE WHATEVER FIELDS ON THE
* RECORD WITH THEIR NEW VALUES AT THIS POINT
REWRITE IN-MASTER-RECORD.
EVALUATE WS-FILE-STATUS
. . .
END-EVALUATE.
Note that READ and DELETE use the file name; WRITE and REWRITE use the record name. DELETE uses the word 'RECORD' after the file name, do not connect it to the file name with a hyphen. REWRITE will not change the number of records in a file, it only alters the data in an existing record. All these statements support the INVALID KEY/NOT INVALID KEY clauses.
The SELECT for indexed files requires the following clauses:
ORGANIZATION IS INDEXED
ACCESS IS DYNAMIC
RECORD KEY IS field-1
FILE STATUS IS working-storage-field-2.
With indexed files, the key is actually part of the record. No two records on the same file can have the same value for this key; the key must uniquely identify the record. working-storage-field-2 is a PIC X(2).
Before a read or write the record key must be populated with a value by the program. When a read is executed the index of the file is searched for the value of the record key. If the record is in the file then a matching entry will be in the index. If a write is executed then the system will ensure that the file doesn't already contain a record for that key.
The value of the file status after the statement is executed will indicate whether or not the statement successfully executed. If not, the value will also indicate what went wrong. This field can be interrogated with an IF or EVALUATE:
SELECT IN-MASTER
ORGANIZATION IS INDEXED
ACCESS IS DYNAMIC
RECORD KEY IS IN-CUSTOMER-NBR
FILE STATUS IS WS-FILE-STATUS.
. . . . . .
MOVE WS-CUST-NO TO IN-CUSTOMER-NBR.
READ IN-MASTER.
EVALUATE WS-FILE-STATUS
WHEN '00'
PERFORM 300-SUCCESSFUL-READ
WHEN '23'
PERFORM 400-RECORD-NOT-FOUND
WHEN '92'
DISPLAY 'FILE NOT OPEN'
WHEN OTHER
DISPLAY 'UNEXPECTED FILE ERROR: ' WS-FILE-STATUS
END-EVALUATE.
See the File Errors Page for a list of file status values and their meanings.
As an alternative the INVALID KEY clause can be used:
MOVE WS-CUST-NO TO IN-CUSTOMER-NBR.
READ IN-MASTER
INVALID KEY
PERFORM 400-RECORD-NOT-FOUND
NOT INVALID KEY
PERFORM 300-SUCCESSFUL-READ
END-READ.
This lumps all unsuccessful operations into one group so this may not be an appropriate method.
Writes work similarly:
SELECT OT-MASTER
ORGANIZATION IS INDEXED
ACCESS IS DYNAMIC
RECORD KEY IS OT-CUSTOMER-NBR
FILE STATUS IS WS-FILE-STATUS.
. . . . . .
MOVE WS-CUST-NO TO OT-CUSTOMER-NBR.
WRITE OT-MASTER-RECORD.
EVALUATE WS-FILE-STATUS
WHEN '00'
ADD +1 TO WS-NBR-RECORDS-WRITTEN
WHEN '22'
PERFORM 500-RECORD-ALREADY-EXISTS
WHEN '92'
DISPLAY 'FILE NOT OPEN'
WHEN OTHER
DISPLAY 'UNEXPECTED FILE ERROR: ' WS-FILE-STATUS
END-EVALUATE.
Writes also support INVALID KEY/NOT INVALID KEY.
Indexed files can be accessed as if they were sequential files. In the SELECT change the ACCESS IS DYNAMIC to ACCESS IS SEQUENTIAL. Can then use READs (include AT END) and WRITEs as if the file was sequential. Do not specify a RECORD KEY or a FILE STATUS.
Indexed files support two other type of operations, but these require the file to be opened for input and output at the same time. Access cannot be sequential. These operations are DELETE and REWRITE.
The DELETE statement will delete the current record. This requires a successful READ before the DELETE.
OPEN I-O IN-MASTER.
. . . . . .
MOVE WS-CUST-NO TO IN-CUSTOMER-NBR.
READ IN-MASTER.
EVALUATE WS-FILE-STATUS
. . .
END-EVALUATE.
DELETE IN-MASTER RECORD.
EVALUATE WS-FILE-STATUS
WHEN '00'
PERFORM 200-SUCCESSFUL-DELETE
WHEN '43'
DISPLAY ' NO CURRENT RECORD. REL KEY = ' WS-RELATIVE-KEY
. . .
END-EVALUATE.
The REWRITE statement will update the current record. This requires a successful READ before the REWRITE.
OPEN I-O IN-MASTER.
. . . . . .
MOVE WS-CUST-NO TO IN-CUSTOMER-NBR.
READ IN-MASTER.
EVALUATE WS-FILE-STATUS
. . .
END-EVALUATE.
* IF THE READ WAS SUCCESSFUL YOU WILL UPDATE WHATEVER FIELDS ON THE
* RECORD WITH THEIR NEW VALUES AT THIS POINT
REWRITE IN-MASTER-RECORD.
EVALUATE WS-FILE-STATUS
. . .
END-EVALUATE.
Note that READ and DELETE use the file name; WRITE and REWRITE use the record name. DELETE uses the word 'RECORD' after the file name, do not connect it to the file name with a hyphen. REWRITE will not change the number of records in a file, it only alters the data in an existing record. All these statements support the INVALID KEY/NOT INVALID KEY clauses.
An indexed file can have more than one key (a multi-key file). There will be one key designated as the primary key, all other keys are secondary, or alternate, keys. Alternate keys do not have to be unique within the file, primary keys do.
The SELECT for such a file still requires the RECORD KEY clause (which identifies the primary key). Immediately following the RECORD KEY clause are any ALTERNATE KEY clauses. Any alternate keys that are used in the program are listed here. If an alternate key does not have to be unique specify it WITH DUPLICATES. All alternate keys must be part of the record.
SELECT IN-RENTAL-CAR-MASTER
ORGANIZATION IS INDEXED
ACCESS IS DYNAMIC
RECORD KEY IS IN-LICENSE
ALTERNATE KEY IS IN-VIN
ALTERNATE KEY IS IN-STYLE WITH DUPLICATES
ALTERNATE KEY IS IN-MFR WITH DUPLICATES
FILE STATUS IS WS-FILE-STATUS.
When the file is opened the access path defaults to that of the primary key. To use an access path of one of the alternate keys, use the START command. You first prime the alternate key field with some value that you want to start with. You then execute a START with either 'EQUALS', 'NOT <' or '>'. For example:
MOVE 'HONDA' TO IN-MFR.
START IN-RENTAL-CAR-MASTER
KEY NOT < IN-MFR
END-START.
This will change the access path to that of the alternate key IN-MFR and then position the file pointer so that it is at the first record whose key is equal to or greater than the specified value.
MOVE SPACES TO IN-VIN.
START IN-RENTAL-CAR-MASTER
KEY > IN-VIN
END-START.
This one changes the access path to that of IN-VIN. The record pointer here is set to the first record in the file with an IN-VIN value higher than that of the specified value.
MOVE 'MIDSIZE' TO IN-STYLE.
START IN-RENTAL-CAR-MASTER
KEY EQUALS IN-STYLE
END-START.
This one changes the access path to that of IN-STYLE. The record pointer here is set to the first record in the file with an IN-STYLE that matches that of the specified value.
The file status needs to be checked after the START, or the INVALID KEY clause used. One cause of an unsuccessful START is if no such record can be found in the file.
The START NOT < and START > do not actually read a record, even if the START was successful in postioning the file pointer.
The READ NEXT statement can then be used to read records along the current access path. The file status must be checked after each READ NEXT or the INVALID KEY clause used. One common cause of an unsuccessful READ NEXT is reaching the end of the file. AT END can also be used.
* GET MANUFACTURER FROM KEYBOARD
DISPLAY 'ENTER CAR MANUFACTURER: ' WITH NO ADVANCING.
ACCEPT WS-MFR.
MOVE WS-MFR TO IN-MFR.
* ATTEMPT TO START FILE AT SPECIFIED MANUFACTURER
START IN-RENTAL-CAR-MASTER
KEY NOT < IN-MFR
INVALID KEY DISPLAY 'NO RECORDS ON FILE FOR ' WS-MFR
NOT INVALID KEY PERFORM 300-READ-NEXT-ON-MFR
END-START.
PERFORM 300-READ-NEXT-ON-MFR UNTIL WS-EOF-SW = 'Y'.
. . .
300-READ-NEXT-ON-MFR.
READ IN-RENTAL-CAR-MASTER NEXT RECORD
AT END MOVE 'Y' TO WS-EOF-SW
NOT AT END PERFORM 310-CHECK-MFR
END-READ.
310-CHECK-MFR.
IF IN-MFR = WS-MFR
DISPLAY 'CAR FOUND FOR MFR ' WS-MFR ' LICENSE ' IN-LICENSE
ELSE
MOVE 'Y' TO WS-EOF-SW
END-IF.
In these examples we will have PROGRAM1 call PROGRAM2. The first example shows a static call. PROGRAM2 is compiled into PROGRAM2.OBJ.
CALL 'PROGRAM2.OBJ'.
PROGRAM2 requires either a 'GOBACK' or 'EXIT PROGRAM' in order for control to return to PROGRAM1. A 'STOP RUN' will terminate all programs.
When a program containing static calls is compiled the object code of all statically called programs is combined with the object code of the program with the calls into one package. The advantage is that the object code of called programs need not be present when the program is run, all object code is included in the one package. The disadvantage with this is that if one of the called programs is changed any program that statically calls it must be re-compiled so that the new version is included.
The other type of call is a dynamic call. When a dynamic call is executed the object code for the called program is searched for on disk (it is not part of the calling program's object code) and then executed. The advantage is that the newest version of the called program is always executed - no need to re-compile the calling program.
To change this to a dynamic call add the following to PROGRAM1's WORKING-STORAGE SECTION. PROGRAM2.OBJ must exist when PROGRAM1 is run.
01 WS-CALLED-PGM PIC X(12) VALUE 'PROGRAM2.OBJ'.
The CALL now looks like:
CALL WS-CALLED-PGM.
Static calls have the called program as a literal in the call, dynamic calls have a field. The value of that field is the name of the called program.
Passing some fields from PROGRAM1 to PROGRAM2 requires a little more work. First, only WORKING-STORAGE fields from PROGRAM1 can be passed to PROGRAM2. These fields are specfied in the CALL as follows:
CALL 'PROGRAM2.OBJ'
USING WS-FIELD-1 WS-FIELD-2 WS-FIELD-3
WS-FIELD-4 WS-FIELD-5
END-CALL.
This can be done with either static or dynamic calls. Group-level fields can be passed so if all 5 of the above fields were under the same group-level field you could just specify it in the CALL.
PROGRAM2 needs to know some fields are coming in. This is done on the PROCEDURE DIVISION statement of PROGRAM2:
PROCEDURE DIVISION USING LK-FIELD-A LK-FIELD-B LK-FIELD-C
LK-FIELD-D LK-FIELD-E.
The fields can have the same name as they do in PROGRAM1. In PROGRAM2 these fields are defined in the LINKAGE SECTION, which is in the DATA DIVISION and follows the WORKING-STORAGE SECTION. The fields in the two USING clauses must appear in the same order. Remember that while the names don't have to match the data types and sizes do.
All fields are passed by reference (meaning that pointers to the fields are passed and not copies of the data) so that PROGRAM2 can alter any of them.
Say that a 3-digit number will be input into your program. Define:
01 WS-INPUT-FIELDS.
05 WS-INPUT-NUMBER PIC ZZ9.
01 WS-NUMERIC-FIELDS.
05 WS-NUMERIC-NUMBER PIC 999.
In the PROCEDURE DIVISION:
DISPLAY 'Enter a number (0-999): ' WITH NO ADVANCING.
ACCEPT WS-INPUT-NUMBER.
MOVE WS-INPUT-NUMBER TO WS-NUMERIC-NUMBER.
* CALCULATIONS ARE TO BE DONE USING WS-NUMERIC-NUMBER