Re: Why is the optimiser choosing a sub-optimal plan?

Stephen Cresswell <pgsql-performance@stephen-cresswell.net> writes:
> I have the a table with two indexes...

(1) Tell us about the other table, mobile_summary_type.

(2) Did you transcribe the second query plan correctly?  I have a hard
time believing that EXPLAIN printed two Index Cond lines for the same
indexscan.

(3) What PG version is this, exactly?

(4) Are you doing anything funny like disabling autovacuum/autoanalyze?
The rowcount estimates in the "good" plan seem rather far away from
reality, and it's not obvious why, particularly here:

>               ->  Seq Scan on mobile_summary_type mst
> (cost=0.00..18.80 rows=880 width=64) (actual time=0.014..0.016 rows=4
> loops=1)

            regards, tom lane