From: | Magnus Hagander <magnus(at)hagander(dot)net> |
---|---|
To: | Stepan Yankevych <Stepan_Yankevych(at)epam(dot)com> |
Cc: | pgsql-hackers <pgsql-hackers(at)postgresql(dot)org> |
Subject: | Re: CDC feature request |
Date: | 2021-03-18 13:07:41 |
Message-ID: | CABUevExjjSe3maOx-0bwF=WEATb-+ibQdRcbw1mX=1SGppB_pw@mail.gmail.com |
Views: | Raw Message | Whole Thread | Download mbox | Resend email |
Thread: | |
Lists: | pgsql-hackers |
On Thu, Mar 18, 2021 at 2:03 PM Stepan Yankevych <Stepan_Yankevych(at)epam(dot)com>
wrote:
> Hi All!
>
>
>
> Hopefully I’m using correct mail list
>
> If not please show me right direction 😊
>
>
>
> I’m quite struggling without native Change Data Capture feature in
> PostgreSQL.
>
>
>
> That would be great to implement it, possibly in not so complicated way.
>
>
>
> Can Logical replication be a little be modified or reused to do not
> replicate data into destination table as is but to insert each change into
> “change table” (like in oracle 11 CDC)?
>
> That change table should have at lease few additional columns
>
> - Operation (I/D/U)
> - txid
> - Commit_time_stamp
>
>
If you look at logical decoding, that's basically what you have, isn't it?
It won't go into a table, but you can consume it into one if you want. Look
at for example wal2json for examples of how to consume it -- but the system
is pluggable so you can build your own or use one of the others available
plugins.
--
Magnus Hagander
Me: https://www.hagander.net/ <http://www.hagander.net/>
Work: https://www.redpill-linpro.com/ <http://www.redpill-linpro.com/>
From | Date | Subject | |
---|---|---|---|
Next Message | Fujii Masao | 2021-03-18 13:12:48 | Re: fdatasync performance problem with large number of DB files |
Previous Message | Amit Langote | 2021-03-18 13:06:33 | Re: A reloption for partitioned tables - parallel_workers |