Re: RFC tool to support development / operations work with slony replicated databases

From: "Andrew Hammond" <andrew(dot)george(dot)hammond(at)gmail(dot)com>
To: pgsql-general(at)postgresql(dot)org
Subject: Re: RFC tool to support development / operations work with slony replicated databases
Date: 2007-03-06 18:49:05
Message-ID: 1173206945.375137.89320@30g2000cwc.googlegroups.com
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-general

On Mar 6, 5:44 am, k(dot)(dot)(dot)(at)secdat(dot)com (Kenneth Downs) wrote:
> Andrew Hammond wrote:
> > Each release will include a directory that has the same name as the
> > full release tag. This directory must contain all the scripts to be
> > applied.
>
> Have you considered using a data dictionary instead,

Data dictionary is a pretty generic term, could you please show me
what you mean by this?

> so that you can
> simply diff the structures and generate DDL to bring a database to its
> current state?

To fully express the semantics of a change, DDL is often interspersed
with DML. A trivial example is refactoring a table to reflect a 1 to n
relationship. In order to do this successfully, you must first execute
DDL to create a new table, then execute DML to pull data from the old
table into the new table, and finally execute DDL to remove the
columns in the old table as necessary.

> If your scripts contain data modification code, they can be preserved,
> but using a dictionary smooths over a lot of the quirkiness of the
> scripts-based approach.

Again, I'd be very interested in seeing what you mean by a dictionary.

Andrew

In response to

Browse pgsql-general by date

  From Date Subject
Next Message Andrej Ricnik-Bay 2007-03-06 19:20:27 Re: No buffer space available
Previous Message Reuven M. Lerner 2007-03-06 18:48:35 Re: Database slowness -- my design, hardware, or both?