Re: [sqlite] Query planner regression with FTS4: slower path is chosen

On 2/9/62 16:57, Paul wrote: I has been a while without response, so I just bumping this message. 19 July 2019, 14:21:27, by "Paul" : I have a test case when the regression can be observed in queries that use JOINs with FTS4 tables, somewhere in between 3.22.0 and 3.23.0. For some reason

Re: [sqlite] Query planner regression with FTS4: slower path is chosen

I has been a while without response, so I just bumping this message. 19 July 2019, 14:21:27, by "Paul" : > I have a test case when the regression can be observed in queries that > use JOINs with FTS4 tables, somewhere in between 3.22.0 and 3.23.0. > For some reason the planner decides to

[sqlite] Query planner regression with FTS4: slower path is chosen

I have a test case when the regression can be observed in queries that use JOINs with FTS4 tables, somewhere in between 3.22.0 and 3.23.0. For some reason the planner decides to search non-FTS table first then scan the whole FTS table. Version 3.22.0 is the last unaffected, while issue is still