Edmund Mergl <E(dot)Mergl(at)bawue(dot)de> writes:
> ... That's the reason I
> was talking about the strange UPDATE behavior of
> PostgreSQL. If it can determine a specific number
> of rows in a reasonable time, it should be able to
> update these rows in the same time frame.
Not necessarily --- this table has a remarkably large number of indexes,
and all of them have to be updated when a tuple is replaced. So the
amount of work is significantly greater than simply finding the tuples
will require.
As I posted later, I think that much of the problem comes from poor
handling of equal-key cases in our btree index routines...
regards, tom lane