IBM Mainframe Forum

Hi,

Could someone help me with this please?

Input:
----+----1----+----2----+----3----+----4----+----5----+----6----+----7--
14 Feb 00:43:51.7967 => ****GI Identifier= 20111111111111111111
14 Feb 00:43:51.7967 => SORTWORKSFORUAAAAAAAAAAAAAAAAccountKey/Output Validation error
14 Feb 00:43:51.7967 => XXXXXXXXXXXXXXXXXXXXXXXXXNo Data XXXXXXXXXXXXXXXXXXXXXXXXXXX
14 Feb 00:43:51.7967 => SIMPLECHARLIE
14 Feb 00:43:52.3124 => ****GI Identifier= 20112222222222222222
14 Feb 00:43:52.3124 => B1234567890AAAAAAAAAAAAAAAAccountKey/Output Validation error
14 Feb 00:43:52.3124 => XXXXXXXXXXXXXXXXXXNo Data XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX
14 Feb 00:43:52.3124 => SIMPLECHARLIE


Expected Output:
----+----1----+----2----+----3----+----4----+----5----+----6----+----7--
20111111111111111111 SORTWORKSFORU
20112222222222222222 B1234567890AA


Important Points:
1. The length of the Input file is 251 and it's a FB file.
2. The number in the input (Eg.. 20111111111111111111 always starts with "2011" and starts from the 44th position.LENGTH = 20
3. The text in the input (Eg.. SORTWORKSFORU) always starts from the 25th position in the line next to the number (2011..).
LENGTH=13

4. The text in the output should start from 22nd position after the number.

Query:

I know I need to use IFTHEN(WHEN=GROUP).
Not sure how to do it?

djprakash1ml,

Use the following DFSORT JCL which will give you the desired results.

Thanks a lot !!

SORT FIELDS=COPY
INREC IFTHEN=(WHEN=GROUP,RECORDS=2,
BEGIN=(25,18,CH,EQ,C'****GI IDENTIFIER='),PUSH=(252:44,20,SEQ=1))
OUTFIL INCLUDE=(272,1,ZD,EQ,2),BUILD=(252,20,X,25,13)

In this case the SOROUT dataset can a record lenght of 34
Is this because we use BUILD?

Any alternatives in this case?

Thanks

djprakash1ml wrote:In this case the SOROUT dataset can a record lenght of 34 Is this because we use BUILD?

Yes that is you showed as the desired output in your original post. Do you want the output to be different?

No this is fine.
But I was having lenght of the output dataset as 80 and the job was abending.
Is there anyway we can avoid the record length restriction?

djprakash1ml,

I am assuming you hard coded the LRECL of 80 in your JCL. You really don't have to code the DCB parameters for a sort job as sort products are capable of calculating them from the input or by the INREC/OUTREC statements. So remove the hard coded values in your JCL. Alternatively if you DO want to create a 80 byte file by padding them with spaces , then change your BUILD statement to the following

Thanks for your patient reply. My sincere thanks.
So dumb of me (I am referring to the BUILD statement change..)

Thanks again,
David