From: | Merlin Moncure <mmoncure(at)gmail(dot)com> |
---|---|
To: | rey <reywang(at)optonline(dot)net> |
Cc: | Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us>, pgsql-general(at)postgresql(dot)org |
Subject: | Re: psql copy command - 1 char limitation on delimiter |
Date: | 2010-09-26 13:05:44 |
Message-ID: | AANLkTikRwGi6kUXzVPfdFYV=j-97XUNeddrGChE1B3-=@mail.gmail.com |
Views: | Raw Message | Whole Thread | Download mbox | Resend email |
Thread: | |
Lists: | pgsql-general |
On Sat, Sep 25, 2010 at 3:12 PM, rey <reywang(at)optonline(dot)net> wrote:
> On 09/25/2010 10:03 AM, Tom Lane wrote:
>>
>> rey<reywang(at)optonline(dot)net> writes:
>>
>>>
>>> Why limit this to a single character?
>>>
>>
>> Performance. Believe it or not, breaking fields at the delimiter is
>> a significant factor in COPY speed.
>>
>> regards, tom lane
>>
>>
>
> True, but just for 5% to 10% degradation here.
> For RDBMS, correct indexes and good logical design we are talking about 10
> times or more performance gains.
>
> Who cares about 10% waste here? Is it Oracle and other commercial RDBMS no
> such limitation.
Believe it or not, data loading performance is one of the most common
standard metrics people used to benchmark databases. A large class of
applications need to slam data in the db as quickly as possible, do
some work, and slam it out/dump it. Copy performance matters.
merlin
From | Date | Subject | |
---|---|---|---|
Next Message | A B | 2010-09-26 13:51:08 | Re: Index on points |
Previous Message | Andrew Hunter | 2010-09-26 12:22:12 | Re: "IN" in a geometric database using data type "point" |