large scale <[EMAIL PROTECTED]> writes: > Aggregate (cost=353859488.21..353859488.21 rows=1 > width=78) > -> Merge Join (cost=1714676.02..351297983.38 > rows=1024601931 width=78) > -> Index Scan using genescript_genomseqid on > genescript (cost=0.00..750.35 rows=25115 width=62) > -> Sort (cost=1714676.02..1714676.02 > rows=6801733 width=16) > -> Seq Scan on mouseblathuman > (cost=0.00..153685.33 rows=6801733 width=16)
That plan seems odd to me too. Have you done VACUUM ANALYZE on these tables? I would think that a hash join would be preferable. You might need to increase the SORT_MEM parameter to let the whole smaller table be stuffed into memory before the planner will think so, though. Try setting it to 10000 or so (ie, 10 MB). regards, tom lane ---------------------------(end of broadcast)--------------------------- TIP 3: if posting/reading through Usenet, please send an appropriate subscribe-nomail command to [EMAIL PROTECTED] so that your message can get through to the mailing list cleanly