Re: [HACKERS] Another nasty cache problem

From: Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us>
To: chris(at)bitmead(dot)com
Cc: pgsql-hackers(at)postgreSQL(dot)org
Subject: Re: [HACKERS] Another nasty cache problem
Date: 2000-02-04 06:33:45
Message-ID: 1812.949646025@sss.pgh.pa.us
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-hackers

Chris Bitmead <chrisb(at)nimrod(dot)itg(dot)telstra(dot)com(dot)au> writes:
>> No ... portals are a backend concept ...

> Since when?

> According to the old doco you do...

> select portal XX * from table_name where ...;

> fetch 20 into XX.

That still works if you spell it in the SQL-approved way,
DECLARE CURSOR followed by FETCH.

> If the PQexec() is called with "fetch 20" at a time
> wouldn't this mean that you wouldn't exhaust front-end
> memory with a big query?

Sure, and that's how you work around the problem. Nonetheless
this requires the user to structure his queries to avoid sucking
up a lot of data in a single query. If the user doesn't have any
particular reason to need random access into a query result, it'd
be nicer to be able to read the result in a streaming fashion
without buffering it anywhere *or* making arbitrary divisions in it.

In any case, psql doesn't (and IMHO shouldn't) convert a SELECT
into a series of FETCHes for you.

regards, tom lane

In response to

Browse pgsql-hackers by date

  From Date Subject
Next Message Thomas Lockhart 2000-02-04 07:47:27 Re: [HACKERS] docs and createlang patch for plperl
Previous Message Tom Lane 2000-02-04 06:06:53 Re: [HACKERS] how to deal with sparse/to-be populated tables