"Kevin Grittner" <Kevin.Grittner@wicourts.gov> writes:
> As I poked around at this, it started to seem familiar. I had
> previously posted about this query's performance under 8.2.4.
> http://archives.postgresql.org/pgsql-performance/2007-10/msg00087.php
Well, that thread gave some of the missing details, such as
Table "public.ControlRecord" Column | Type | Modifiers
--------------------+------------------------+-----------countyNo | "CountyNoT" | not null
but what the heck is "CountyNoT"?
It looks like 8.3 is failing to realize that it can propagate the
countyNo = 53 condition down to this table's scan, as 8.2 did;
but there's not enough details here to guess why not.
regards, tom lane