MVSFORUMS.com

[ramy2016]

Good day everyone and Happy Holiday Season!

Could someone, please, let me know how to accomplish the following with DFSORT step (SORT or ICETOOL).

I have a dataset, FB LRECL=80, where there could be duplicate Account-No records accompanied by Account-Class values.
There could be three types of duplicates:
- duplicate Account-No recs within Account-Class '9'
- duplicate Account-No recs between Account-Class '9' and Account-Class '5'
- duplicate Account-No recs within Account-Class '9' AND a record from Account-Class '5'
- there could never be duplicate Account-No recs within Account-Class '5'.

All duplicates within Account-Class '9' should go to an Output. Also recs of Account-Class '9' take precedence over Class '5', so if there are duplicate Accounts from both Classes, only Account-Class '9' should make it to the Output.
And non-dups from Account-Class '5' should also be in the Output.

Input records could be sorted in any order, as needed.

Input.

[kolusu]

ramy2016,

You need to use the trick of joinkeys matching the file to it self. The second file will count the number of duplicates and upon matching every record has the counter. Now we can just eliminate the records with dup count > 1 and class of '5' on OUTFIL

Here is an untested DFSORT JCL which will give you the desired results