Re: Persist MVCC forever - retain history

On 05.07.2020 08:48, Mitar wrote:
> Hi!
>
> On Fri, Jul 3, 2020 at 12:29 AM Konstantin Knizhnik
> <k(dot)knizhnik(at)postgrespro(dot)ru> wrote:
>> Did you read this thread:
>> https://www.postgresql.org/message-id/flat/78aadf6b-86d4-21b9-9c2a-51f1efb8a499%40postgrespro.ru
>> I have proposed a patch for supporting time travel (AS OF) queries.
>> But I didn't fill a big interest to it from community.
> Oh, you went much further than me in this thinking. Awesome!
>
> I am surprised that you are saying you didn't feel big interest. My
> reading of the thread is the opposite, that there was quite some
> interest, but that there are technical challenges to overcome. So you
> gave up on that work?
No, I have not gave up.
But...
There are well known problems of proposed approach:
1. Not supporting schema changes
2. Not compatible with DROP/TRUNCATE
3. Presence of large number of aborted transaction can slow down data
access.
4. Semantic of join of tables with different timestamp is obscure.

I do not know how to address this issues. I am not sure how critical all
this issues are and do them made this approach unusable.
Also there is quite common opinion that time travel should be don at
application level and we do not need to support it at database kernel level.

I will be glad to continue work in this direction if there is some
interest to this topic and somebody is going to try/review this feature.
It is very difficult to find some motivation for developing new features
if you are absolutely sure that it will be never accepted by community.