RE: Question on buffer lengths vs SQL reported lengths

I suspect that if you created a table with 100 date columns, and then created a loop which initialized 1,000,000 rows into that table, you would end up with an object size closer to 4 byte columns than 10 byte columns. Not forgetting the additional overhead stored with a table object like descriptions, etc.

But let's assume your problem is that you are trying to create a function like the undocumented view qsys2.journal_view (which you may want to look at) and you need the 10 byte offset, not the 4 byte offset. I believe that SYSCOLUMNS is built off of QADBIFLD. However that uses the 4 byte definition also.
Good luck, or try the journal_view function.

call qsys2.display_journal_entry_info(2, --not sure what 2 means
'journal_library',
'journal name',
Sequence_number,
'receiver library',
'receiver name',
'journal code (aka PT, DL, etc)';

Rob Berendt