Use top level join instead of subquery in hold queue position query
This is more readily optimized by the Postgres planer.
Note also, for very large data sets (lots of holds, on the order of 100k+ active), the following is also advised:
ALTER TABLE action.hold_copy_map alter column target_copy SET statistics 500, alter column hold set statistics 500;
ANALYZE action.hold_copy_map;
This gives the planner better data about the hold-copy-map n-distinct.
git-svn-id: svn://svn.open-ils.org/ILS/trunk@20221
dcc99617-32d9-48b4-a31d-
7c20da2025e4