Message-ID: <4CE21D4A.4070200@gmx.net>
Date: Tue, 16 Nov 2010 06:57:30 +0100
From: Andreas <maps.on@gmx.net>
User-Agent: Mozilla/5.0 (Windows; U; Windows NT 5.1; de;
	rv:1.9.2.12) Gecko/20101027 Thunderbird/3.1.6
MIME-Version: 1.0
To: pgsql-sql@postgresql.org
Subject: multi table import from 1 denormalized source table
Content-Type: text/plain; charset=UTF-8; format=flowed
Content-Transfer-Encoding: 7bit

Hi,

I frequently get tables from spreadsheets to import into the DB.

Usually it looks like this:
A1, A2, A3, A4,   B1, B2, B3,   with optional C1, C2,   D1, D2, ...

and there is a 1:n relation between A and B.
If provieded the C would be 1:1 to A and D 1:1 to B.

Up until now I let a VBA script order the source table by A, then scan 
the table line by line and create a new entry in the target table A* and 
fetch its serial ID everytime the script figures that A changed.
With this IDa create 1 C* and as many B*s until A changes again ... and 
of course fetch IDb to attach the D* records with a foreign key column.

Now I'm trying to get away w/o the VBA stuff.

Is there a clever way to split such denormalized sources while still 
obtaining the needed IDs to connect everything?