Amit Langote wrote:
> On 2018/04/18 7:11, Alvaro Herrera wrote:
>
> @@ -1546,8 +1546,8 @@ match_clause_to_partition_key(RelOptInfo *rel,
> case PARTITION_STRATEGY_HASH:
> cmpfn = get_opfamily_proc(part_scheme->partopfamily[partkeyidx],
> - op_righttype, op_righttype,
> - HASHEXTENDED_PROC);
> + part_scheme->partopcintype[partkeyidx],
> + op_righttype, HASHEXTENDED_PROC);
>
> This change is not quite right, because it disables pruning. The above
> returns InvalidOid as there are no hash AM procedures (in pg_amproc) whose
> lefttype and righttype don't match.
Makes sense. Still, I was expecting that pruning of hash partitioning
would also work for pseudotypes, yet it doesn't.
--
Álvaro Herrera https://www.2ndQuadrant.com/
PostgreSQL Development, 24x7 Support, Remote DBA, Training & Services