Starting with version 9.6, PostgreSQL supports parallel queries. This support for parallelism has been improved gradually over time, and version 11 has added even more functionality to this important feature. In this section, we will take a look at how parallelism works and what can be done to speed things up.
Before digging into the details, it is necessary to create some sample data, as follows:
test=# CREATE TABLE t_parallel AS
SELECT * FROM generate_series(1, 25000000) AS id;
SELECT 25000000
After loading the initial data, we can run our first parallel query. A simple count will show what a parallel query looks like in general:
test=# explain SELECT count(*) FROM t_parallel;
QUERY PLAN
-------------------------------------------------------------------------------------------------
Finalize Aggregate (cost=241829.17..241829.18 rows=1 width=8)
-> Gather (cost=241828.96..241829.17 rows=2 width=8)
Workers...