as you see, the first 8 bytes is the primary account and all accounts to the right are secondary accounts. I want to have a list of primary and secondary accounts with the primary accounts on one column and the secondary accounts on the second column. the primary account repeats as the secondary account for each row.
What this does is build one 18 byte record for each possible secondary account (including the first) with the primary account in pos 1-8 and the secondary account in pos 11-18. There will obviously be many, many records where the secondary account will be blank (equal to spaces). The OUTFIL OMIT process eliminates those records that contain a space where the secondary account would start.
Unfortunately, you are correct - after taking another look at the manual, I see that it does say that the n/ parameter can only be used on the OUTREC statement within an OUTFIL statement. I missed noting that restriction earlier. Sorry about that.
If you can live with doing it in TWO steps, however, you can do the INREC processing as OUTFIL OUTREC processing in step 1 and the OMIT processing in step 2.