Re: GiST "choose subtree" support function to inline penalty

+1,

This sounds reasonable.  Authors of existing GiST opclasses wouldn't have trouble to keep compatible with new PostgreSQL versions.

I see one more use case for "choose subtree" instead "penalty".  When R*-tree chooses subtree, it considers to only extension of selected bounding box, but also overlap increase of bounding boxes.  This strategy should have a positive effect of tree quality, besides I never seen it has been measured separately.  It probably kind of possible to implement using "penalty" method assuming you have reference to the page in GISTENTRY.  But that doesn't seems a correct way to use the GiST interface.  Additionally, you don't know the attribute number to get the correct key in multicolumn indexes.  Having "choose subtree" method will make it possible to implement this strategy in correct way.  However, this use case is kind of opposite to Darafei's one, because it should make choosing subtree slower (but better).

The Russian Postgres Company