Tom Lane <tgl@sss.pgh.pa.us> writes:
> Nope, that is certainly not the explanation, because the hash table is
> loaded in the (single) call of the Hash node at the start of the query.
> It is static all through the sampled-and-not executions of the Hash Join
> node, which is where our problem is.
At the risk of asking a stupid question, it's not perchance including that
hash build in the first sample it takes of the hash join node?
--
greg