4 ms·

Other databases have optimizer hints for that, but apparently there’s resistance to adding hints into Postgres [1]. I find hints very useful (in Oracle world) f

Other databases have optimizer hints for that, but apparently there’s resistance to adding hints into Postgres [1]. I find hints very useful (in Oracle world) for quickly working around some optimizer glitch (and figure out the long term solution later). There actually is a “pg_hint_plan” [2] Postgres module that adds some hints to Postgres optimizer (but it’s not part of Postgres core as I understand): [1] https://wiki.postgresql.org/wiki/OptimizerHintsDiscussion https://wiki.postgresql.org/wiki/OptimizerHintsDiscussion [2] https://github.com/ossc-db/pg_hint_plan https://github.com/ossc-db/pg_hint_plan

smackeyacky 5y ago
I always use the rule of thumb that if you need optimizer hints for your query, its time to fix the query or the indexes. Of course in the teeth of an outage that generally isn't an option. The tools on SQL server for query optimization are better then postgres, but that doesn't always mean problems like the OP don't still occur. I have had more than one SQL server query plan go to hell over time as the tables grew and the optimizer started to make odd decisions. Edit: to make this a little clearer, I regard the SQL server query optimizer system as "lawful evil". It follows its own rules relentlessly.
- tanelpoder 5y ago
  Yep, when the app code and the schema/partitioning/index design works _with_ the database as intended, not _against_ the database engine’s intended use, a whole class of SQL performance and plan stability problems tends to go away. Yep, agreed, hints should be temporary workarounds (or just experimentation aid when developing/testing SQL performance to validate some hypothesis). But many of these temporary workarounds end up being permanent temporary workarounds :-) That’s why Oracle now has hints and parameters to make the optimizer ignore all other hints in the SQL code :-)
  - smackeyacky 5y ago
    Working out which hints are operating must be difficult.
- j16sdiz 5y ago
  For a cost-based-optimizer to work, the cost tuneables must be good. These are very difficult to get right in practice.
- dap 5y ago
  How would you fix the query or indexes here? The CTE seems clearly a hack, effectively functioning as a hint. It doesn’t seem like it counts as fixing the query. I’m skeptical of this rule of thumb. I’m sure people misuse hints. But if someone thinks they “need” hints, my guess is they’ve tried changing the query and indexes.

smackeyacky 5y ago

I always use the rule of thumb that if you need optimizer hints for your query, its time to fix the query or the indexes. Of course in the teeth of an outage that generally isn't an option. The tools on SQL server for query optimization are better then postgres, but that doesn't always mean problems like the OP don't still occur. I have had more than one SQL server query plan go to hell over time as the tables grew and the optimizer started to make odd decisions. Edit: to make this a little clearer, I regard the SQL server query optimizer system as "lawful evil". It follows its own rules relentlessly.

tanelpoder 5y ago
Yep, when the app code and the schema/partitioning/index design works _with_ the database as intended, not _against_ the database engine’s intended use, a whole class of SQL performance and plan stability problems tends to go away. Yep, agreed, hints should be temporary workarounds (or just experimentation aid when developing/testing SQL performance to validate some hypothesis). But many of these temporary workarounds end up being permanent temporary workarounds :-) That’s why Oracle now has hints and parameters to make the optimizer ignore all other hints in the SQL code :-)
- smackeyacky 5y ago
  Working out which hints are operating must be difficult.
j16sdiz 5y ago
For a cost-based-optimizer to work, the cost tuneables must be good. These are very difficult to get right in practice.
dap 5y ago
How would you fix the query or indexes here? The CTE seems clearly a hack, effectively functioning as a hint. It doesn’t seem like it counts as fixing the query. I’m skeptical of this rule of thumb. I’m sure people misuse hints. But if someone thinks they “need” hints, my guess is they’ve tried changing the query and indexes.