ossc-db / pg_plan_advsr Goto Github PK

Depending on the gcc version, a warning will occur when executing the make command.
ex. gcc (GCC) 7.3.1

priority:low

pg_plan_advsr.c:1854:5: warning: this ‘else’ clause does not guard... [-Wmisleading-indentation]
     else
     ^~~~
pg_plan_advsr.c:1860:2: note: ...this statement, but the latter is misleadingly indented as if it were guarded by the ‘else’
  appendStringInfo(str, "host=%s port=%s dbname=%s user=%s", pghost, pgport, dbName, login);
  ^~~~~~~~~~~~~~~~

After some discussion over here ankane/dexter#32 I'v following suggestion:

Optimization assumes that columns are independent, so a query plan can go wrong if there are functional dependencies (FD) between columns which he doesn't know. Real world datasets often have FDs between columns because of denormalization or just because the data is like this.

Actually, finding FDs for all columns could become a rather hard problem to solve. But hopefully this challenge becomes feasible if applied only to columns as part of given slow queries.

Expected behaviour:

PostgreSQL offers CREATE STATISTICS in order to inform the planner about FDs. So, looking e.g. at the freely available database "ATP tennis tour", and given a slow query involving tourney_id, tourney_name, pg_plan_advsr could make a specific suggestion like this:

create statistics atp_matches_2019_tourney_id_tourney_name (dependencies) 
	on tourney_id, tourney_name 
	from atp_matches_2019;

Hi Team - I will be thankful you guys if i get to know that if i can use pg_plan_advsr in window OS. I am waiting for prompt reply.

Hi,

Note: For now, This extension is in POC phase. Not production ready.

Note says, its not ready for Production. When it will be marked for Production?

Thanks