Re: NOLOGGING option, or ?

From: Bruce Momjian <pgman(at)candle(dot)pha(dot)pa(dot)us>
To: Luke Lonergan <llonergan(at)greenplum(dot)com>
Cc: Alon Goldshuv <agoldshuv(at)greenplum(dot)com>, pgsql-hackers(at)postgresql(dot)org
Subject: Re: NOLOGGING option, or ?
Date: 2005-06-01 23:00:34
Message-ID: 200506012300.j51N0Yp21000@candle.pha.pa.us
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-hackers

Luke Lonergan wrote:
> Bruce,
>
> > The problem with a new command is that it becomes unclear when you
> > should use COPY and when LOAD DATA, and it confuses users, and has
> > maintenance overhead. If Bizgres wants a new command name, go for it,
> > but it is unlikely that the community release is going to go in that
> > direction, unless there is a fundamental agreement that COPY is broken
> > and needs a major revamp, and I have heard no talk of that.
>
> The question of whether COPY should be improved or whether the changes
> should take the form of a new command is separate from the question of
> whether the performance of the load path in PostgreSQL needs improvement.
>
> The 90% performance increase (from 12 MB/s to 21 MB/s) that Alon reported
> comes from replacing the parsing logic within COPY. I believe that the
> parsing logic in COPY is fundamentally broken from a performance
> perspective, and may be broken from a functionality perspective WRT embedded
> backslashes.

COPY works as designed. The idea that some guy we have never heard of
is going to appear and rewrite COPY's processing and tell us that the
existing code is actually broken seems pretty arrogant to me. If it is
broken (meaning doesn't work as designed), please show us facts rather
than conjecture.

Oh, and the "Our COPY improvements are so fundamental that they deserve
a new command name" also has a similar flavor.

(Please explain how you handle literal delimiters and nulls with no
escape processing.)

> One of the reasons to consider a LOAD DATA command is that we can isolate
> the need for performance improvements and special syntax from the concerns
> of preserving the legacy behavior of COPY for use as the primary mechanism
> for DUMP and RESTORE.

This seems like a case where GreenPlum's priorities and the community's
priorities might not match. There is much more work required on your
part if you are going to convince the community it needs a new data
loading command, and starting out with the assumption in emails that it
is going to be a newly named command isn't the best approach. That is
my fundamental point.

--
Bruce Momjian | http://candle.pha.pa.us
pgman(at)candle(dot)pha(dot)pa(dot)us | (610) 359-1001
+ If your life is a hard drive, | 13 Roberts Road
+ Christ can be your backup. | Newtown Square, Pennsylvania 19073

In response to

Responses

Browse pgsql-hackers by date

  From Date Subject
Next Message Bruce Momjian 2005-06-01 23:02:17 Re: NOLOGGING option, or ?
Previous Message Simon Riggs 2005-06-01 22:47:17 Re: NOLOGGING option, or ?