Message-ID: <463FE405.40704@autoledgers.com.au>
Date: Tue, 08 May 2007 10:44:21 +0800
From: Paul Lambert <paul.lambert@autoledgers.com.au>
User-Agent: Thunderbird 1.5.0.10 (Windows/20070221)
MIME-Version: 1.0
To: pgsql-sql@postgresql.org
Subject: Re: Query RE using COPY
References: <463FD7B8.9030700@autoledgers.com.au>
	<36e682920705071916l27d64129j2d3c9d296fbbd6e5@mail.gmail.com>
In-Reply-To: <36e682920705071916l27d64129j2d3c9d296fbbd6e5@mail.gmail.com>
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Content-Transfer-Encoding: 7bit

Jonah H. Harris wrote:
> COPY mytbl (col1, col2, col3, col4, col6) FROM 'myfile.txt' WITH
> DELIMITER AS '^' QUOTE '\f' CSV HEADER;
> 
> 

I would rather not do it this way as I use the same load script at all 
customer sites where the extracts and requirements may vary. I.e. one 
customer may not use the location field mentioned above, but another might.

What I would prefer to do, if possible, is set that column in the 
database to 'not required' so that if I get an extract that doesn't have 
that column in the file, the copy doesn't care.

As it stands now I generally have to drop all the unneeded columns from 
  numerous tables, perform my load and then re-add them back again.

When we're talking about 40 or more tables in each database some tables 
with several hundred columns... your suggestion would be a bit 
cumbersome - particularly if it is only one or two columns in each table 
that that the client doesn't need.

-- 
Paul Lambert
Database Administrator
AutoLedgers