Hello Experts,
I'm currently tuning a massive SQLscript procedure and looking for pointers from people who have done similar exercises before.
What I noticed in the current procedure is,
There are a couple of internal tables,
which is then used for some intermediate processing - which is again stored in an internal table.
The var_out is then generated by joining a few of these intermediate internal tables.
I feel this is forcing the procedure to run sequentially(wait till all the intermediate results are materialised before building resultset),
and thought of constructing a massive select with all intermediate selects as sub-queries.
My PO is not in favor of this idea as this decreases readability of code ( which kinda makes sense ).
Then i started exploring for ways to achieve this without changing code ( i.e have temp tables in procedure but make the
compiling/execution happen by constructing a massive select statement ).
I found some hints which looked like they can help but,
also could not find any examples or documentation for them.
NO_CS_ITAB_IN_SUBQUERY - Prefer/Avoid column engine itab in subquery.
Can someone share some insights ,
1. If this is the right direction to go?
2. How do we do we achieve this keeping the readability of code intact?
Thanks,
Vinay