Re: [PERFORM] sequence scan problem

[John Beaver]

<chuckle> You're right - for some reason I was looking at the (18 rows) at the bottom. Pilot error indeed - I'll have to figure out what's going on with my data. Thanks! Tom Lane wrote: John Beaver <[EMAIL PROTECTED]> writes: Ok, here's the explain analyze result. Again, this is Postgres 8.3.3 and I vacuumed-analyzed both tables directly after they were created. Merge Join (cost=1399203593.41..6702491234.74 rows=352770803726 width=22) (actual time=6370194.467..22991303.434 rows=15610535128 loops=1) ^^^^^^^^^^^ Weren't you saying that only 50 rows should be returned? I'm thinking the real problem here is pilot error: you missed out a needed join condition or something. SQL will happily execute underconstrained queries ... regards, tom lane