Re: increasing effective_cache_size slows down join queries by a factor of 4000x - Mailing list pgsql-general

From Michael Lewis
Subject Re: increasing effective_cache_size slows down join queries by a factor of 4000x
Date
Msg-id CAHOFxGojXjwzZTkFBqQtxJOTJKkD8Qnu38PqGp2dm-nTN6mMZw@mail.gmail.com
Whole thread Raw
In response to Re: increasing effective_cache_size slows down join queries by a factor of 4000x  (A Shaposhnikov <artyom@gmail.com>)
List pgsql-general
I think that theoretically if the planner sees a condition like

a.column1 > constant_value, and it also has a condition like a.column1 = b.column2 then it could autogenerate the b.column2 > constant_value condition. And of course > could be <, <=, >= and <>

But I could be wrong, particularly with respect to nulls and how = is mostly the same as 'is not distinct from' except for 'select null = null' being null and 'select null is not distinct from null' being true. Perhaps if either/both columns have a not null constraint, then this could still be done. Whether or not it should be a separate question.

pgsql-general by date:

Previous
From: A Shaposhnikov
Date:
Subject: Re: increasing effective_cache_size slows down join queries by a factor of 4000x
Next
From: Michael Lewis
Date:
Subject: Re: Rows From but with Subqueries (or a cleaner non-array-using alternative)?