4 ms·
I both like and am annoyed by this. I like it because they actually do legitimately treat it as a bug and are quite responsive in publishing fixes - I'm annoye
by munk-a 7y ago
I both like and am annoyed by this. I like it because they actually do legitimately treat it as a bug and are quite responsive in publishing fixes - I'm annoyed by this because their scale of responsiveness is (for very valid reasons including measuring how a proposed fix will effect other cases and the cost of relying on mostly volunteer labour) quite a bit behind the scale of time I need a fix in.
Using CTEs you are still able to hack the optimizer and while in pg12 they're making that fence optional it is remaining an option, that seems to be their primary conceit to needing to bludgeon an underperformant query - though, per their policy, queries that remain underperformant are getting rarer as they fix up the planner more and more.