I have several times now run into what seems
like similar performance problems with some
of my postgresql queries.
I have a view that runs reasonably quicky.
I use this view in a subselect in another
query and that query too runs reasonably
quicky.
The view returns some unwanted duplicate
rows so I modify it using either DISTINCT
or GROUP BY to eliminate them.
View still runs reasonably quickly.
I use the modified view as a subselect as
above, but now the query runs 2-3 orders
of magnitude more slowly than before.
Before I go through the effort of putting
together a specific and concise test case,
has anyone seen this general pattern and
have an explanation or advice? (PG-8.2.3)