On Thu, Jan 5, 2012 at 5:04 AM, Benedikt Grundmann
<bgrundmann@janestreet.com> wrote:
> We are experiencing a big performance regression in some queries.
> In those cases the planner seems to choose a nested loop index
> scan instead of hashing the index once and then joining.
I think you probably need to post EXPLAIN ANALYZE output from the
actual queries to get useful advice, probably to pgsql-performance,
rather than here.
It's hard to believe that enable_nestloop=off is doing anything other
than masking whatever the real problem is, but it's hard to tell what
that problem is based on the information provided.
--
Robert Haas
EnterpriseDB: http://www.enterprisedb.com
The Enterprise PostgreSQL Company