IBM Mainframe Forum

(Initially posted to the DFSort forum in error, then moved it out here as we use Syncsort - )

Hello,

I have an input file as follows (File A) and another VSAM lookup file (File B)

The rules -

The keys from B will need to be looked up in input A in multiple columns. If there is a match of the keys in any of the columns of input , it will be written to a
matched file, and the unmatched records from A will be written to another file. In the example below, the columns in file A to be searched are 1:11, 19:29, 31:41, 43:53 , and 55:65.

File B (key file) - (cols 1:11)



Input file A


Output file (unmatched)


Output file (matched)


Is there any elegant way I could use splice/joinkeys for the multiple fields without having to go through several passes with joinkeys?
Thank you!

I don't have SyncSort documentation. Does SyncSort have something equivalent to the Slash (/) operator that DFSORT has for OUTFIL (OUTREC and BUILD)?

If so, you can process your VSAM file, creating the keys. Sort them dropping the duplicates.

Then the JOIN becomes simple.

Hi BillyBoyo,

Thank you.. I forgot to mention that the VSAM file B has already beens sorted and contains unique keys.
Beyond this point is where I am stumped - would the only option be multiple Join steps for the columns in A ?

Thank you!

What do you mean about the unique keys on the VSAM file? Records one and two show repeated key values. Is there something which is regarded as unique on the VSAM file? What order is it sorted in?