On Apr 21, 2008, at 10:44 PM, Christophe wrote:
> I'd like a tool that would automatically create these scripts, and I
> wondered if anything like this existed. The theory would be that it
> would consider two databases a and b, and produce the appropriate
> script to change b's schema to match a.
> Does anything like this exist? If not, I might have a new project...
However it happens, you need to *capture* the schema changes that need
to be applied to each database. I don't think you can just compare
schemas and guarantee that database A will be transformed in exactly
the same way as database B. For example, suppose the last text column
name in database B was different from A. How can you determine if the
column was renamed or if the column was dropped and a new column was
added? The semantics of that difference could be very important.
It would be nice if PostgreSQL had some kind of unique reference for
the column, but I think columns are just numbered sequentially as they
are added. It would also be neat to have a built-in way to log the
John DeSoi, Ph.D.
In response to
pgsql-general by date
|Next:||From: David Wilson||Date: 2008-04-22 20:31:42|
|Subject: Rapidly decaying performance repopulating a large table|
|Previous:||From: Joris Dobbelsteen||Date: 2008-04-22 19:42:32|
|Subject: Re: table as log (multiple writers and readers)|