Re: [PERFORM] Chaotic query planning ?

2017-02-01 Thread Philippe Ivaldi
Albe Laurenz wrote > […] > Experiment with raising join_collapse_limit and from_collapse_limit to 11. Thank you, this solve the problem. > Alternatively, optimize the join order by hand and don't tune the parameters. What is surprising is that there is no apparent/logical optimal strategy. Bes

[PERFORM] Chaotic query planning ?

2017-01-19 Thread Philippe Ivaldi
Hi, I'm new in this mailing list, sorry if my post is not well formed. First of all, I would thank all the team and the contributors around PostgreSQL for their work. My question… The explain analyze of the following code is https://explain.depesz.com/s/VhOv ✂⋯✂⋯✂⋯✂⋯✂⋯✂⋯✂⋯✂⋯✂⋯✂⋯✂⋯✂⋯✂⋯✂⋯✂⋯✂⋯ WIT