I see what you mean, now. Thanks!

- DAP 

-----Original Message-----
From: Dan Wright [mailto:wright at smx.pair.com] 
Sent: Thursday, July 22, 2004 9:57 PM
To: David Parker
Cc: slony1-general at gborg.postgresql.org
Subject: RE: [Slony1-general] partitioning with replication


Maybe I mis-understood your original request.  But here is an example of
what I was talking about..

Suppose you have a table called "people", and you want all of the male
rows to replicate onto one server, and all of the female rows to
replicate onto another server.

One way to do this would be to use inheritance.  Make two tables:
"male", and "female" that inherit from "people".  Any time you insert a
row into people, you use a trigger or other mechanism to really insert
it into the male or female table.

Now, you have your people split out, but "select * from people" still
does the right thing.  Once you have that, you can set up slony to
replicate the male table onto one server and the female table onto
another server.

Apologies if that isn't what you were asking for.

Daniel J. Wright                           wright at pair.com
Lead Software Developer, pairNIC   https://www.pairnic.com
pair Networks, Inc.                    http://www.pair.com

On Thu, 22 Jul 2004, David Parker wrote:

> How does table inheritance help me to populate tables on another host?

> I have, say, one master and 3 slaves, all with the same schema, but 
> table A in each slave as a sub-set of the data in table A on the
master.
>  I was thinking maybe of defining a view, A, across tables like 
> host1_A, host2_A, etc., then defining a rule to map updates to the 
> appropriate table, then allow replication to work from that point on. 
> But I still have the question of naming on each host (though I guess I

> could use the same view scheme on each slave, as well).
>  I'm pretty new to postgres, so there may well be an aspect of table 
> inheritance of which I'm completely ignorant. Did you have a 
> particular approach in mind?
>
> Thanks!
>
> - DAP
>
> 	-----Original Message-----
> 	From: Dan Wright [mailto:wright at smx.pair.com]
> 	Sent: Thu 7/22/2004 6:15 PM
> 	To: David Parker
> 	Cc: slony1-general at gborg.postgresql.org
> 	Subject: Re: [Slony1-general] partitioning with replication
>
>
>
>
> 	It sounds like you are doing the kind of thing that table
inheritance is
> 	supposed to solve.  If you use inheritance in your DB tables,
you could get
> 	exactly what you are asking for from slony right now.
>
> 	-Dan
>
> 	Daniel J. Wright                           wright at pair.com
> 	Lead Software Developer, pairNIC   https://www.pairnic.com
> 	pair Networks, Inc.                    http://www.pair.com
>
> 	On Thu, 22 Jul 2004, David Parker wrote:
>
> 	> In one area of our application, we would like to be able to
capture an
> 	> update to a master table, and select the target replication
table based
> 	> on some key value in the update data, e.g. "if key_field == A,
replicate
> 	> to table X on host A; if key_field == B, replicate to table X
on host
> 	> B", etc.
>
>
>