IBM Mainframe Forum

Hi,

I am trying to select the matching records from the vsam file based on the key.
Its numeric of 12 length. When I run this program...it does not return any error. But for some reasons, its not able to match any records from the VSAM file (but there are matching records).

key in the vsam file:



It displays in the log as
example:


Could you please advise what could be problem?



Code'd

That output is not from that program. The output shows 12 bytes of key, the key is 15 bytes.

ah, sorry... I missed to enter them

121212121212121NOT FOUND
141212121212121NOT FOUND
121312121212121NOT FOUND

Hi Billy,

I just got the original copybook and found that the VSAM key are:
S code - s9(12) comp-3
Ac no - s9(16) comp-3
ser no- s9(1) comp-3

I think since the comparing key of VSAM is in packed decimal, and so its not matching with the numeric key in FILE1.

1. Is it possible to compare the complete key of the VSAM together or do I need to compare each field separately? Please suggest.
2. Could you please advise how do I compare a numeric field with a Packed decimal?

I think you first need to dump a few records from the VSAM data set and verify the key. S9(12) COMP-3 will require 7 bytes, not 6, to store and S9(16) COMP-3 will require 9 bytes, not 8, to store. So instead of having 3 fields of 6, 8, 1 bytes for a total key length of 15 you have 7, 9, 1 bytes for a total key length of 17. There is something wrong since you are sure the key length is 15. I suspect the key fields are actually S9(11) COMP-3 and S9(15) COMP-3 which would have lengths of 6 and 8 bytes, respectively.

In any case, you will need to compare each field separately since they are packed decimal. Assuming you do have 6, 8, 1 bytes of packed decimal you will need to define variables as W 6 P, W 8 P, and W 1 P for the packed decimal values you want to compare your key to.

Your question 2 makes no sense -- a packed decimal value IS numeric so comparing a numeric field to a packed decimal field is not really a question; it's more the way the system works. If you use the P modifier on your variables, they will be packed decimal and you won't have to think about the numeric to packed decimal comparison.

That worked. Thanks Robert and Billy Boyo.