Re: bitmap scans, btree scans, and tid order

Tom Lane wrote:
> "Jeffrey W. Baker" <jwbaker@acm.org> writes:
> 
>>I see that Tom has already done the infrastructure work by adding
>>getmulti, but getmulti isn't used by nodeIndexscan.c, only
>>nodeBitmapIndexscan.c.  Will btree index scans be executed by creating
>>in-memory bitmaps in 8.1, or will some scans still be executed the usual
>>way?
> 
> 
> We aren't going to remove the existing indexscan behavior, because
> bitmap scans lose the ordering of the underlying index.  There are many
> situations where that ordering is important.  (See for instance the
> recent changes to make MAX/MIN use that behavior.)

Would you take a patch that retained the optimized executions of plans 
returning 1 tuple and also fixed the random heap problem?

-jwb