Re: Strange execution plan

2021-07-08 Thread Manuel Weitzman
> On 08-07-2021, at 17:13, Manuel Weitzman wrote: > > Option A: Use a common table expression to "force" the usage of > test_json_data_idx > >WITH json_matching_rows AS ( >SELECT t.* >FROM test ti >WHERE t.json_data @> '{"book":{"title":"In Search of Lost Time"}}'

Re: Strange execution plan

2021-07-08 Thread Manuel Weitzman
> On 08-07-2021, at 17:13, Manuel Weitzman wrote: > > I'd recommend you using cursors for pagination in general (I know it > may not be possible for you, just wanted to explain as it could be > useful). By the way, I mean cursor pagination as the general concept. I'm not talking about Postgre

Re: Strange execution plan

2021-07-08 Thread Manuel Weitzman
> On 08-07-2021, at 04:26, Joel Frid wrote: > > Hi, > > On my production environment (PostgreSQL 13.3), one of my queries runs very > slow, about 2 minutes. > I noticed that it does not use an execution plan that I anticapited it would. > > The query is : > > SELECT t.* > FROM test t > W

Strange execution plan

2021-07-08 Thread Joel Frid
Hi, On my production environment (PostgreSQL 13.3), one of my queries runs very slow, about 2 minutes. I noticed that it does not use an execution plan that I anticapited it would. The query is : SELECT t.* FROM test t  WHERE  t."existe" IS true and  t.json_data" @> '{"book":{"title"

Strange execution plan

2021-07-08 Thread Joel Frid
Hi, On my production environment (PostgreSQL 13.3), one of my queries runs very slow, about 2 minutes. I noticed that it does not use an execution plan that I anticapited it would. The query is : SELECT t.* FROM test t  WHERE  t."existe" IS true and  t.json_data" @> '{"book":{"title"