> can we see an explain analyze at least?
>
Hi,
Well, it won't be necessary - I mean it looks just like the explain I
sent in my first post.
BUT I found the real cause of my problem - the "fk2" field from my
example had not only an index, but it was also a foreign key to
another table.
I believe the update took so long because pgsql was checking if the
changes don't break the referential integrity.
When I dropped the FK constraint (and index too - just in case) the
update took around 3 minutes which is acceptable.
So - problem solved, postgres good. ;) But isn't there a way to make
some bulk operations without having to drop indexes/FKs?
Something that would work like:
begin transaction + forget about RI
make some lenghty operation (update/delete...)
if RI is OK then commit; else rollback
Thanks,
MS