I’ve recently been reminded, in no uncertain terms, of the importance of running the “ANALYZE” (or “VACUUM ANALYZE”) command periodically on a PostgreSQL database installation.

My problem:
Amongst other tables, a production database I administer has three tables: two moderately large tables, A⁽¹⁾ and B, and one quite large one (>700 000 records), C (for “Cross”), which is essentially the full cross product of A and B, with an additional piece of information for each tuple (record). So I’m associating some information with every combination of records from tables A and B.
Continue reading “ANALYZE your PostgreSQL databases!” →