Re: BUG #15904: ERROR: argument of LIMIT must not contain variables - Mailing list pgsql-bugs

From Lakradi Marwan
Subject Re: BUG #15904: ERROR: argument of LIMIT must not contain variables
Date
Msg-id CAF5OE-zgnmn++jAVX1WVxsL+yROjvJp-kQop8DT=2g8MZLhtvA@mail.gmail.com
Whole thread Raw
In response to Re: BUG #15904: ERROR: argument of LIMIT must not contain variables  ("David G. Johnston" <david.g.johnston@gmail.com>)
List pgsql-bugs
Thank you for the clarification and for the time you gave me.

I was convinced of that kind of answer, I still wanted to try, on a misunderstanding you never know ! 
I will find a workaround, however the LIMIT_TILL clause would have been welcomed in this particular case.

Wishing you an excellent weekend.
Marwan L.



Le jeu. 11 juil. 2019 à 18:45, David G. Johnston <david.g.johnston@gmail.com> a écrit :
On Thu, Jul 11, 2019 at 9:26 AM Lakradi Marwan <lakradimarwan@gmail.com> wrote:
Thank you for your quick feedback,

In my opinion, your query should return :
   Id, Text
1| 1, 'one'

and be interpreted as due to variable type :
LIMIT CASE WHEN 1 = 1 THEN 1 ELSE 2 END;

Am I wrong in the way I see the situation ?

I think that in the case of a variable, the data should be retrieved until the condition is reached. Like LIMIT_TILL {condition}

That's not an unreasonable expectation.  But that isn't how LIMIT is defined.  LIMIT provides a query result max record count to return to the client - mostly to facilitate pagination when used in concert with ORDER BY (for determinism) and OFFSET (to skip already seen records).  That max is a constant determined at plan time which means it cannot rely upon any of the data the query itself may generate.

SQL is set oriented and your expectation is incompatible with that fundamental property of the system.  If you don't want records "after" something you need to apply a inequality filter in the WHERE clause to remove the undesired records.

David J.

pgsql-bugs by date:

Previous
From: David Rowley
Date:
Subject: Re: PG11 - Multiple Key Range Partition
Next
From: Michael Paquier
Date:
Subject: Re: BUG #15905: FATAL: the database system is starting up