In v14 (which is currently in beta), autoanalyze will process the partitioned
table automatically:
https://www.postgresql.org/docs/14/release-14.html
|Autovacuum now analyzes partitioned tables (Yuzuko Hosoya, Álvaro Herrera)
|Insert, update, and delete tuple counts from partitions are now propaga
On Thu, Jul 22, 2021 at 01:32:51PM +0200, Kamil Frydel wrote:
> table_1 and table_2 are hash partitioned using volume_id column. Usually we
> make analyze on partitions. We do not make analyze on the partitioned table
> (parent).
> However, if we run 'analyze' on the partitioned table then planner
Hi,
we faced a performance issue when joining 2 partitioned tables
(declarative partitioning). The planner chooses nested loop while we
expect hash join.
The query and the plan are available here: https://explain.depesz.com/s/23r9
table_1 and table_2 are hash partitioned using volume_id colu
