Using parallel query

PostgreSQL 9.6 introduces parallel query, after many years of development.

Response times from long running queries can be improved by the use of parallel processing. The concept is that we divide a large task up into multiple smaller pieces. We get the answer faster, but we use more resources to do that.

Very short queries won't get faster by using parallel query, so you'll gain more by thinking about better indexing strategies.

How to do it...

Take a query that needs to do a big chunk of work, such as this:

\timing
SELECT count(*) FROM accounts;
count
---------
1000000
(1 row)
Time: 261.652 ms
SET max_parallel_workers_per_gather = 8;
SELECT count(*) FROM accounts;
count
---------
1000000
(1 row)
Time: 180...