Stephan Szabo wrote:
>
>
>IIRC, he was. I think the thing causing the difference between his times
>and the ones we saw typically when doing the tests was that he didn't have
>an index on the fktable's referencing column.
>
>
A common mistake, can't count how often I created this one... And not
easy to find, because EXPLAIN won't explain triggers.
I'm planning to create some kind of fk index wizard in pgAdmin3, which
finds out about fks using columns that aren't covered by an appropriate
index. Maybe this check could be performed (as a NOTICE) when the fk is
created?
Regards,
Andreas