Andrew Nesheret wrote:
Richard Huxton wrote:
Richard Huxton wrote:
I'm putting together a small test case to see if I can reproduce your
behaviour here.
Does the attached small script misbehave in the same way as your real
data? From here it works fine when the fkey is ON ... RESTRICT.
I'm right in thinking that your "nodes" fkey is RESTRICT on update and
delete?
You script is worked fine on same database, where is difference?
Good question. There *must* be a difference. Why can't either of us see it?
1. Try adding another 5 million rows to the test "traffic" table and see
if that makes any difference. It shouldn't.
2. Run a "vacuum verbose sf_ipv4traffic" and see if there's a lot of
dead rows. I shouldn't have thought there are.
We might want to move this to the performance mailing-list, and see if
anyone has an idea over there.
--
Richard Huxton
Archonet Ltd
---------------------------(end of broadcast)---------------------------
TIP 9: In versions below 8.0, the planner will ignore your desire to
choose an index scan if your joining column's datatypes do not
match
