Re: [PERFORM] Troubleshooting query performance issues - resolved (sort of)

2013-09-27 Thread Jim Garrison
We have traced this to the *addition* of a two-column index. > -Original Message- > From: pgsql-performance-ow...@postgresql.org [mailto:pgsql- > performance-ow...@postgresql.org] On Behalf Of Jim Garrison > Sent: Wednesday, September 25, 2013 8:58 AM > To: pgsql-performance@postgresql.org

Re: [PERFORM] Troubleshooting query performance issues - Resolved (sort of)

2013-09-27 Thread Jim Garrison
We have traced this to the *addition* of a two-column index. The two tables in question both have single-column indexes on two foreign keys, say columns A and B. The query joins the two large tables on A and B. With only the two indexes, the query plan does a bitmap AND on the index scan re

Re: [PERFORM] Troubleshooting query performance issues

2013-09-25 Thread bricklen
On Wed, Sep 25, 2013 at 8:58 AM, Jim Garrison wrote: > I spent about a week optimizing a query in our performance-testing > environment, which has hardware similar to production. > > I was able to refactor the query and reduce the runtime from hours to > about 40 seconds, through the use of CTEs