How do you grab a string in COBOL from a file when the position is unknown?

Question

I'm new to the site as well as COBOL. I am trying to write a program that reads in an 80 byte file, and finds a certain string and grabs another string that is positioned right after that. The only issue I'm having with this is that the starting position of the string is not always in the same byte throughout the file. For example, the string I am trying to find below is the LENGTH(#####) string that appears twice throughout the file:

LENGTH(14909135) FILEID(DD:EDIREC) MSGDATE(130723) MSGDATELONG(20130723)
MSGTIME(091053) MSGSEQO(001390) MSGNAME(00008557) MSGSEQNO(00001)
SESSIONKEY(XXXXXXXX) DELIMITED(E) SYSNAME(XXXXX-XX) SYSLEVEL(XXXX) TIMEZONE(L)
DATATYPE(E) EDITYPE(XXX) SENDERFILE(#####) RECFM(????) RECLEN(#) RECDLM(E)
UNIQUEID(XXXXXXXX) SYSTYPE(##) SYSVER(#);
RECEIVED ACCOUNT(XXXX) USERID(XXXXXXXX) CLASS(#E2) CHARGE(3) LENGTH(14911043)
FILEID(DD:EDIREC) MSGDATE(130723) MSGDATELONG(20130723) MSGTIME(093045)
MSGSEQO(001392) MSGSEQNO(00000) SESSIONKEY(XXXXXXXX) DELIMITED(C)
SYSNAME(XXXXX-XX) SYSLEVEL(XXXX) TIMEZONE(L) DATATYPE(E) EDITYPE(UNFORMATTED)
SENDERFILE(XXXXXXXXXXXXX) RECFM(????) RECLEN(0) RECDLM(C) UNIQUEID(XXXXXXXX)
SYSTYPE(24) SYSVER(5);

Notice the two LENGTH(#####) strings. The below code manages to count the amount of times the length string appears as well as grab the final length string count (what I really want, the numbers within the length string), but only when they are in these two positions:

    WORKING-STORAGE SECTION.

    01 WS-INPUT-RECORD   PIC X(80).

    01 WS-STRINGS.

       05 LENGTH-STRING      PIC X(7) VALUE 'LENGTH('.

    01 WS-COUNTERS.

       05 WS-MSG-COUNT  PIC 9(11).

    01 WS-CHAR-TOTALS.

       05 CHAR-TOTAL  PIC 9(11) VALUE ZEROS.

       05 TMP-TOTAL  PIC X(11) VALUE ZEROS.

    ......

    PROCEDURE DIVISION.

    2200-GET-MSG-TOTAL.

        INSPECT WS-INPUT-RECORD
        TALLYING WS-MSG-COUNT FOR ALL LENGTH-STRING.

    2300-CHAR-TOTAL.

        IF WS-INPUT-RECORD(1:7) = LENGTH-STRING

           MOVE WS-INPUT-RECORD(8:9) TO TMP-TOTAL

           UNSTRING TMP-TOTAL DELIMITED BY ')'
           INTO CHAR-TOTAL

        END-IF

        IF WS-INPUT-RECORD(61:7) = LENGTH-STRING

           MOVE WS-INPUT-RECORD(68:9) TO TMP-TOTAL

           UNSTRING TMP-TOTAL DELIMITED BY ')'
               INTO CHAR-TOTAL

        END-IF

The code works great for the two positions shown in the example input above. But it won't work if LENGTH(####) ends up in any other byte position. Other than coding 80 IF statements to check for every byte in the file for the string, is there an easier way to go about getting those values inside of the length parens? I've checked a lot of other posts and I've thought about using pointers or tables but I can't quite seem to figure it out.

Answer 1

You can use a "perform varying" loop to look at each block of the string within each line, where each block is a string the length of the string you are looking for. Here is an example that works in OpenCobol:

   IDENTIFICATION DIVISION.
   PROGRAM-ID. FIND-STRING.

   ENVIRONMENT DIVISION.
   INPUT-OUTPUT SECTION.
   FILE-CONTROL.
   SELECT IN-FILE ASSIGN TO 'SAMPLE-LEN.TXT'
       ORGANIZATION IS LINE SEQUENTIAL.

   DATA DIVISION.
   FILE SECTION.
   FD  IN-FILE.
   01  IN-RECORD                        PIC X(80).

   WORKING-STORAGE SECTION.
   01  END-OF-FILE-SWITCH               PIC XXX VALUE 'NO '.
       88  END-OF-FILE                  VALUE 'YES'.
   01  STRING-MARKER                    PIC X(7) VALUE 'LENGTH('.
   01  STRING-MARKER-LENGTH             PIC 99 VALUE 7.
   01  STRING-SOUGHT                    PIC X(11).
   01  STRING-INDEX                     PIC 99.
   01  RECORD-LENGTH                    PIC 99 VALUE 80.

   PROCEDURE DIVISION.
   MAIN.
       OPEN INPUT IN-FILE
       PERFORM UNTIL END-OF-FILE
           READ IN-FILE
               AT END
                   SET END-OF-FILE TO TRUE
               NOT AT END
                   PERFORM FIND-STRING
           END-READ
       END-PERFORM
       CLOSE IN-FILE
       STOP RUN
       .

   FIND-STRING.
       PERFORM VARYING STRING-INDEX FROM 1 BY 1
           UNTIL STRING-INDEX > (RECORD-LENGTH
                                 - STRING-MARKER-LENGTH)
           IF IN-RECORD(STRING-INDEX:STRING-MARKER-LENGTH) =
              STRING-MARKER
              UNSTRING IN-RECORD(STRING-INDEX
                                 + STRING-MARKER-LENGTH : 10)
                  DELIMITED BY ')' INTO STRING-SOUGHT 
              END-UNSTRING 
              DISPLAY STRING-SOUGHT END-DISPLAY 
           END-IF 
       END-PERFORM 
       . 

Answer 2

Use INSPECT to establish that LENGTH( is on the current record.

Only if present, do the following:

UNSTRING using LENGTH( as a delimiter with two receiving fields.

UNSTRING second receiving field delimited by ) leaving you with the number.

For example:

01  delimiting-field PIC X(7) VALUE "LENGTH(".
01  desitnation-field-1 PIC X.
01  destination-field-2 PIC X(18) JUST RIGHT.

UNSTRING source-field DELIMITED BY delimiting-field INTO desitnation-field-1
                                                         destination-field-2

Abandon destination-field-1. Use destination-field-2 for input to the second UNSTRING.

Use meaningful names, rather than those I have shown to illuminate the example.

So,

    01  WS-INPUT-RECORD                     PIC X(80). 
    01  NUMBER-OF-LENGTHS            BINARY PIC 9(4). 
    01  DELIMITER-COUNT              BINARY PIC 9(4). 
        88  NO-DELIMITERS                   VALUE ZERO. 
        88  ONE-DELIMITER                   VALUE 1. 
    01  LENGTH-OPEN-PAREN                   PIC X(7) 
                                            VALUE "LENGTH(". 
    01  DATA-TO-IGNORE                      PIC X. 
    01  DATA-WITH-LENGTH-VALUE              PIC X(80). 
    01  CLOSING-PAREN                       PIC X VALUE ")". 
    01  VALUE-OF-LENGTH-AN                  PIC X(18) JUST RIGHT.

   THE-STUFF. 
       SET NO-DELIMITERS            TO TRUE 
       INSPECT WS-INPUT-RECORD      TALLYING DELIMITER-COUNT 
                                     FOR ALL LENGTH-OPEN-PAREN
       EVALUATE TRUE 
           WHEN NO-DELIMITERS 
               CONTINUE 
           WHEN ONE-DELIMITER 
               PERFORM              GET-THE-DATA 
           WHEN OTHER 
               PERFORM              OH-DEAR-MORE-THAN-ONE 
       END-EVALUATE 
       . 
   GET-THE-DATA. 
       UNSTRING WS-INPUT-RECORD     DELIMITED BY 
                                    LENGTH-OPEN-PAREN 
           INTO                     DATA-TO-IGNORE 
                                    DATA-WITH-LENGTH-VALUE 
       UNSTRING DATA-WITH-LENGTH-VALUE 
                                    DELIMITED BY CLOSING-PAREN
           INTO                     VALUE-OF-LENGTH-AN 
       DISPLAY "THIS IS WHAT WE FOUND" 
       DISPLAY ">" 
               VALUE-OF-LENGTH-AN 
               "<" 
       . 
   OH-DEAR-MORE-THAN-ONE. 
       DISPLAY "THE FOLLOWING LINE HAS MORE THAN ONE LENGTH(" 
       DISPLAY ">" 
               WS-INPUT-RECORD 
               "<" 
       . 

The technique with the INSPECT to see if the "string" is present can be applied to the other solution accepted so that only if the line contains the value desired is it "searched".

Answer 3

Based on Bill Woodger's comments, here is a better solution. Thank's Bill, for teaching me not to slouch :) I still like looping through each record as a way to catch multiple matches on one line, so I kept that part.

   IDENTIFICATION DIVISION.
   PROGRAM-ID. FIND-STRING-2.

   ENVIRONMENT DIVISION.
   INPUT-OUTPUT SECTION.
   FILE-CONTROL.
   SELECT IN-FILE ASSIGN TO 'SAMPLE-LEN.TXT'
       ORGANIZATION IS LINE SEQUENTIAL
       FILE STATUS IS IN-FILE-STATUS.

   DATA DIVISION.
   FILE SECTION.
   FD  IN-FILE.
   01  IN-RECORD                        PIC X(80).

   WORKING-STORAGE SECTION.
   01  IN-FILE-STATUS                   PIC XX.
   01  END-OF-FILE-SWITCH               PIC XXX VALUE 'NO '.
       88  END-OF-FILE                  VALUE 'YES'.
   01  STRING-MARKER-LEFT               PIC X(7) VALUE 'LENGTH('.
   01  STRING-MARKER-RIGHT              PIC X VALUE ')'.
   01  STRING-MARKER-LENGTH             PIC 99 USAGE BINARY.
   01  STRING-INDEX                     PIC 99 USAGE BINARY.
   01  START-INDEX                      PIC 99 USAGE BINARY.
   01  END-INDEX                        PIC 99 USAGE BINARY.
   01  RECORD-LENGTH                    PIC 99 USAGE BINARY.
   01  SEARCH-LENGTH                    PIC 99 USAGE BINARY.
   01  IS-END-FOUND                     PIC XXX VALUE 'NO '.
       88  END-FOUND                    VALUE 'YES'.
       88  END-NOT-FOUND                VALUE 'NO '.

   PROCEDURE DIVISION.
   MAIN.
       OPEN INPUT IN-FILE

       IF IN-FILE-STATUS NOT = '00'
           DISPLAY 'FILE READ ERROR ' IN-FILE-STATUS
           END-DISPLAY
           PERFORM EXIT-PROGRAM
       END-IF

       PERFORM INITIALIZE-LENGTHS

       PERFORM UNTIL END-OF-FILE
           READ IN-FILE
               AT END 
                   SET END-OF-FILE TO TRUE
               NOT AT END
                   PERFORM FIND-STRING
           END-READ
       END-PERFORM
       PERFORM EXIT-PROGRAM
       .

   INITIALIZE-LENGTHS.
       MOVE FUNCTION LENGTH(IN-RECORD) TO RECORD-LENGTH 
       COMPUTE STRING-MARKER-LENGTH = FUNCTION LENGTH(
           STRING-MARKER-LEFT)
       END-COMPUTE
       COMPUTE SEARCH-LENGTH = RECORD-LENGTH - STRING-MARKER-LENGTH
       END-COMPUTE
       .

   FIND-STRING.
       PERFORM VARYING STRING-INDEX FROM 1 BY 1
           UNTIL STRING-INDEX > SEARCH-LENGTH 
           IF IN-RECORD(STRING-INDEX:STRING-MARKER-LENGTH) =
              STRING-MARKER-LEFT
              COMPUTE START-INDEX = STRING-INDEX 
                  + STRING-MARKER-LENGTH
              END-COMPUTE
              SET END-NOT-FOUND TO TRUE
              PERFORM VARYING END-INDEX FROM START-INDEX BY 1 
              UNTIL END-INDEX > RECORD-LENGTH OR END-FOUND
                  IF IN-RECORD(END-INDEX:
                  FUNCTION LENGTH(STRING-MARKER-RIGHT)) = 
                  STRING-MARKER-RIGHT
                      SET END-FOUND TO TRUE
                  END-IF
              END-PERFORM
              COMPUTE END-INDEX = END-INDEX - START-INDEX - 1
              END-COMPUTE
              DISPLAY IN-RECORD(START-INDEX:END-INDEX)
              END-DISPLAY
           END-IF
       END-PERFORM
       .

   EXIT-PROGRAM.
       CLOSE IN-FILE
       STOP RUN
       .