IBM Mainframe Forum

by **lostcontinent** » Wed Dec 11, 2013 2:36 am

hi there,
I have a DS with content like this:

Key1 Key2 value with leading sign (blank or '-')
0001 1111 -000100
0001 1111 000100
0002 1111 000200
0001 1111 000200
0002 1111 -000300
0002 1111 000300
0001 1111 -000100
0001 1111 000100

and the result should be:

Select all

0002 1111 000200
0001 1111 000200

so I want to get rid off all of the records which have exactly the same keys and value (except for the sign) - a kind of +/- pairs,
so these records have to be removed

Select all

0001 1111 -000100
0001 1111 000100

but these records

Select all

0001 1111 -000200
0001 1111 000100

have to be kept both.

I guess a normal SUM will not work and I first have to split the DS and rejoin, but I really have no idea how to do this.

I hope I explained it clearly enough, sorry for my bad english.

frank

by **NicC** » Wed Dec 11, 2013 7:13 pm

Your data does not match between your examples. there is no

Select all

0001 1111 -000200

in your input and surely you should be dropping 2 lots of

Select all

0001 1111 -000100
0001 1111 000100

not just 1 lot?
and should you not include

Select all

0002 1111 -000300
0002 1111 000300

in the records to be dropped?

What is the LRECL and RECFM of all files?

by **lostcontinent** » Thu Dec 12, 2013 12:43 am

Your data does not match between your examples

Nic,
thank you for the hint.
Yes, these were two different examples. I did not point that out clearly.

The DS has a LRECL of 50, FB.
I just want to get rif of all pairs which have the same key and value except for the sign of the value.

More examples:
1st example:

Select all

0001 1111 -000100
0001 1111 000100
0001 1111 -000100
0001 1111 000100

should result in an empty DS.

2nd Example

Select all

0001 1111 -000100
0001 1111 000100
0001 1111 000100

should result in

Select all

0001 1111 000100

3rd example:

Select all

0001 1111 000100
0001 1111 000100
0001 1111 -000111
0001 1111 -000100

should result in