Re: Design database schemas around a common subset of objects

On 11/7/21 16:14, Erik Wienhold wrote:
> Hi Adrian,
>
>> On 07/11/2021 18:38 Adrian Klaver <adrian(dot)klaver(at)aklaver(dot)com> wrote:
>>
>> My questions for this are:

> I'm thinking about something like merges in Git. Database schemas for
> project-a and project-b are like separate repositories and changes in core
> are merged as necessary. This along with project-specific changes ("o" in
> the diagram below) should result in a sequence of migrations applied to the
> respective databases.
>
> project-a oo--o--ooo-o---o--oo---->
> ___/ _/ _________/
> / / /
> core o--oo-o---o-o--o-->
> \ \__ \
> \ \ \
> project-b o--ooo--o-o-o--ooo----->
>
> These merges (always from core to the projects) form a directed acyclic
> graph from which the migration sequence can be generated using a topological
> ordering.
>
>> 4) Are you using or have you looked at Sqitch(https://sqitch.org/)?
>
> Already looking into it after stumbling across it during research on the
> mailing lists ;)

I think of Sqitch as version control(of database) inside the version
control(VCS e.g. Git). Pretty sure it will cover your use case. Since
you mentioned mailing list I will point you at the Sqitch one
(https://groups.google.com/g/sqitch-users)
>
> Erik
>

--
Adrian Klaver
adrian(dot)klaver(at)aklaver(dot)com