Drew wrote:
> I have a planner question about a multi-join query, where if I rewrite
> the query excluding the table that is left joined to the other tables my
> query speed increases by 75%.
> Here's the plan for the first query (using 8.2b3)
>
-----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
>
> HashAggregate (cost=3223.17..3223.41 rows=19 width=18) (actual
> time=34.913..34.978 rows=122 loops=1)
> And here's the plan for the second query without the first LEFT OUTER
> join (using 8.2b3)
>
------------------------------------------------------------------------------------------------------------------------------------------------------------------------
>
> HashAggregate (cost=8927.87..8936.89 rows=721 width=18) (actual
> time=39.484..39.551 rows=122 loops=1)
Query1: 34.9ms
Query2: 39.5ms
I'm not seeing a huge difference here. Are these the right EXPLAINs?
-- Richard Huxton Archonet Ltd