Re: int8, primary key, seq scan - Mailing list pgsql-general

From Jeff Amiel
Subject Re: int8, primary key, seq scan
Date
Msg-id 4126170A.4030204@istreamimaging.com
Whole thread Raw
In response to int8, primary key, seq scan  (Jeff Amiel <jamiel@istreamimaging.com>)
List pgsql-general
fyi....
outside the function (from psql) I receive these results (which I guess
are expected, because without casting, I assume it assumes an int4)


select  * from draft where check_id=43
and it results in a seq scan...
but if I :
select  * from draft where check_id=int8(43)
it results in an index scan using the primary key.

Jeff


Tom Lane wrote:

>Jeff Amiel <jamiel@istreamimaging.com> writes:
>
>
>>CREATE OR REPLACE FUNCTION public.insert_draft(int4,  numeric)
>>  RETURNS draft AS
>>'DECLARE
>>  passed_customer ALIAS for $1;
>>  passed_amount ALIAS for $2;
>>  draftid int8;
>>  draftrow draft%rowtype;
>>BEGIN
>>insert into draft( amount, customer) values  (passed_amount,
>>passed_customer);
>>select into draftid currval(\'public.draft_check_id_seq\');
>>select into draftrow * from draft where check_id=int8(draftid);
>>return draftrow;
>>END;'
>>  LANGUAGE 'plpgsql' VOLATILE;
>>
>>
>
>Hmm.  I put this function into 7.4.5, minus the explicit cast to int8,
>and it worked just fine (I used a debugger to verify directly that the
>last SELECT was producing an indexscan plan).  I don't see anything in
>the CVS logs to suggest that there was a related bug fix between 7.4.2
>and 7.4.5.  So I'm baffled why it didn't work for you.  Ideas anyone?
>
>            regards, tom lane
>
>---------------------------(end of broadcast)---------------------------
>TIP 9: the planner will ignore your desire to choose an index scan if your
>      joining column's datatypes do not match
>
>
>
>
>


pgsql-general by date:

Previous
From: Wayne Fang
Date:
Subject: libpq on the server
Next
From: the inquirer
Date:
Subject: Authentication problems