COBOL read/store in table

cobol

The goal of this exercise is to read and store an input file into a table then validate certain fields within the input and output any error records. I need to read and store each policy group so that there are just 5 records stored in the table at a time instead of the entire file.

So I need to read in a policy group which is 5 records, do the processing, then read the next 5 records, etc until the end of the file..

This is the input file.
10A 011111          2005062520060625                                    
20A 011111000861038                                                     
32A 011111                            79372                             
60A 0111112020                                             6          4 
94A 011111     080 1                                                    
10A 02222          2005082520060825                                    
20A 022221000187062                                                     
32A 022221                            05038                             
60A 0222212003                                             6          4 
94A 022221     090 1                                                    
....

I was able to load the first 5 records into a table by having my table OCCUR 5 TIMES but I don't know how I would continue that. My code is below. (I wrote it just to see if it was working correctly, but it prints the header line with the first 4 records, instead of just the first 5)
01 TABLES.

05  T1-RECORD-TABLE.                                
    10  T1-ENTRY                OCCURS 5 TIMES      
                                INDEXED BY T1-INDEX.

        15  RECORD-TYPE-10      PIC X(80).          
        15  RECORD-TYPE-20      PIC X(80).          
        15  RECORD-TYPE-32      PIC X(80).          
        15  RECORD-TYPE-60      PIC X(80).          
        15  RECORD-TYPE-94      PIC X(80).          

 copy trnrec10.
 COPY TRNREC20.
 COPY TRNREC32.
 COPY TRNREC60.
 COPY TRNREC94.
.....

Z200-READ-FILES.                               
    READ DISK-IN INTO T1-ENTRY(T1-INDEX)       
        AT END MOVE 'YES' TO END-OF-FILE-SW.   

    WRITE PRINT-RECORD FROM T1-ENTRY(T1-INDEX).

I don't want a step by step for this (though that'd be nice :P) bc I know WHAT I need to do I just don't know HOW to do it bc my textbook and course note are useless to me. I've been stuck on this for a while and nothing I try works.

Best Solution

I'm assuming that every policy group has exactly 5 records with the 5 record types.

You can set up your working storage like this.

05  T1-RECORD.
    10  T1-RECORD-TYPE               PIC XX.
    10  FILLER                       PIC X(78).      

 COPY TRNREC10.
 COPY TRNREC20.
 COPY TRNREC32.
 COPY TRNREC60.
 COPY TRNREC94.

Then your read paragraph would look like this. I assumed that TRNREC10-RECORD was the 01 level of the TRNREC10 copybook. if not, substitute the actual 01 levels in the following code.

2200-READ-FILE.                               
    READ DISK-IN INTO T1-RECORD      
        AT END MOVE 'YES' TO END-OF-FILE-SW.  

    IF END-OF-FILE-SW = 'NO'
        IF T1-RECORD-TYPE = '10'
            MOVE T1-RECORD TO TRNREC10-RECORD
        END-IF
        IF T1-RECORD-TYPE = '20'
            MOVE T1-RECORD TO TRNREC20-RECORD
        END-IF
        ...
    END-IF.

Your write paragraph would look like this

2400-WRITE-FILE.
    WRITE PRINT-RECORD FROM TRNREC10-RECORD
    WRITE PRINT-RECORD FROM TRNREC20-RECORD
    ...

Your processing paragraphs would access the data in the copybook records.

Related Solutions

COBOL Confusion

It sounds like your primary concern is: how does the compiler know how much to allocate in the array if the size changes at run-time?

The answer is that it allocates the maximum amount of space (enough for 90 entries). Note that this is for space in working storage. When the record is written to a file, only the relevant portion is written.

An example:

01  TABLE-SIZE  PIC 9
01  TABLE OCCURS 1 TO 9 TIMES DEPENDING ON TABLE-SIZE
    03 FLD1  PIC X(4)

This will allocate 36 characters (9 multiplied by 4) for TABLE in working storage. If TABLE-SIZE is set to 2 when the record is written to a file, only 8 characters of TABLE will be written (over and above the characters written for TABLE-SIZE, of course).

So, for example, if the memory occupied by TABLE was AaaaBbbbCcccDdddEeeeFfffGgggHhhhIiii, the date written to the file may be the shortened (including size): 2AaaaBbbb.

Similarly, when the record is read back in, both TABLE-SIZE and the relevant bits of TABLE will be populated from the file (setting only the size and first two elements).

I don't believe that the unused TABLE entries are initialised to anything when that occurs. It's best to assume not anyway, and populate them explicitly if you need to add another item to the table.

For efficiency, you may want to consider setting the TABLE-SIZE to USAGE IS COMP.

Java – Reading COBOL datastructures from Java

You could look at JRecord or cb2java. Both allow you to access COBOL files, but neither will generate the full classes.

Update Jan 2011

Since the original answer:

JRecord continues be developed. There is now a JRecord Code generator available as either a standalone program or in the Recordeditor. This Code Generator will build JRecord JRecord code from a COBOL Copybook. See RecordEditor Jrecord CodeGen)
Development on cb2java has stopped
cobol2j has been written. There have been no updates for a year.
There is also Legstar again nothing published for a few years

Update Aug 2017

The RecordEditor has a Generate option for generating Java / JRecord code from a COBOL Copybook. See RecordEditor Code Generation notes for details.

Update Jan 2018

There is some information on generating Java~JRecord code in this question / answer:

How do you generate java~jrecord code for a Cobol copybook