I thought I'd be clever, and instead of checking for the existence of a row
in a table, I simply INSERT the row anyway and wait for an error.

What I didn't expect was that I would get the error condition two seconds
later!

I'm using IB 6.0 on Linux and Interclient 1.6. When I issue the offending
INSERT, two seconds goes by before the SQLException comes back. The table
I'm referencing has 7 columns, a Primary key consisting of two columns, 6
foreign key constraints and seven million rows. The database has been
backed-up/restored less than a week ago, so it shouldn't be disorganised.

Does anybody have any clues as to why the insertion failure should take so
long?
--
Adrian