Scott Marlowe <smarlowe@xxxxxxxxxxxxxxxxx> writes: > So, do you see any obvious, low hanging fruit here? It would help if we were being told the whole truth about the settings being used. The first few plans are clearly suffering from the "enable_seqscan = off" error, but the last few don't seem to be. I don't trust the SHOW ALL at all since it disagrees with the immediately following retail SHOWs --- there is seemingly a whole lot of parameter changing going on that we are not being told about. It'd also be a good idea to know something about the datatypes involved, particularly for the join keys. regards, tom lane
