Optimize query for information_schema.constraint_column_usage

The way the old query was written prevented some join optimizations
because the join conditions were hidden inside a CASE expression.  With
a large number of constraints, the query became unreasonably slow.  The
new query performs much better.

From: Alexey Bashtanov <bashtanov@imap.cc>
Reviewed-by: Ashutosh Bapat <ashutosh.bapat@enterprisedb.com>