Ashutosh Bapat <ashutosh.bapat@enterprisedb.com> writes:
> The reason being, it doesn't look into the subqueries (in FROM clause) to
> infer that p.product_id is essentially product.product_id which is a
> primary key.
Right.
> Attached find a crude patch to infer the same by traversing subqueries.
I think this is probably a bad idea. We could spend an infinite amount
of code this way, with ever-increasing runtime cost and ever-decreasing
usefulness, and where would we stop? I'm also a bit concerned about
allowing illegal queries due to bugs of omission in the
ever-more-complex checking code, which could be quite hard to find, and
when we did find them we'd be faced with a backwards compatibility break
if we fix them.
A larger point is that the patch as proposed doesn't fix the stated
problem, because it only descends into written-out subqueries. It
would only succeed at looking into views if we applied it after
rewriting, rather than in the parser. That's really not going to work.
It would be a complete disaster if the dependencies of a query that
references a view depend on the view's contents.
regards, tom lane