Re: Howto use fields "by name"

["Buck Calabro" <buck.calabro@xxxxxxxxxxxx>]

Hi Antonio,

> I know I could just run the SQL statement. In
> fact it comes from an ODBC connection which is
> behaving too slow under some circunstances,
> and we got some changes, store it in the IFS
> as a set of statements, copy them to a plain,
> flat file, in a file, and now I'd like to run
> them "on the fly" directly. If I use dynamic
> SQL results, I fear, will be as slow (or even
> worse) they were when executed directly under
> ODBC...

Respectfully, this approach will probably not help you with better
performance.  You need to find out why the SQL is running slowly, and
randomly tinkering is not the most efficient way to work that out.

May I suggest that you do STRDBG, then run your SQL and examine the
job log.  When in debug mode, SQL will put hints in the job log about
the decisions that the optimiser was making.  It is very possible that
your ODBC statements need a new index in order to speed them up.

The point being that no matter how you manage to execute those SQL
statements, the exact same SQL engine and optimiser will be working
with them.

In fact, I'd do exactly as Charles suggested and try them as dynamic
statements using PREPARE.  It will be somewhat easier to run
performance tools and benchmarking if you eliminate the network
latency of ODBC.
  --buck