Re: Can an SQL index be defined over multiple tables in a JOIN?

Hello Alan,

Am 04.04.2023 um 22:52 schrieb Alan Cassidy <cfuture@xxxxxxxxxxx>:

Any suggestions on what to look for first to return results more quickly?

It's hard to give hard advice without some more information about the underlying tables and their "geometry" (width, length).

I know that it's really annoying to wait for the population of a subfile.

Is it possible to define an SQL INDEX with JOIN's of up to four or five files?

I don't see any reason why this should not work.

I have a couple of other ideas based on the data in the particular files in the VIEW but wondered if there are any general ideas. Mostly whether an INDEX with JOINs would improve the performance.

My limited knowledge is that each field to be JOINed need to be indexed for maximum performance.

I suggest to maybe think into an entirely different direction. If the database can't deliver the result of a complex SQL statement quickly enough, it might be better to asynchronously calculate that complex SQL in a generic way into an intermediate result table. Is this feasible in your particular scenario?

I have encountered situations (not with IBM i, though) where extensive database redesign was necessary because the "naturally grown" structure wasn't up to the task anymore.

:wq! PoC