Обсуждение: Temporal tables as part of main release
I am looking to use the temporal tables feature for keeping track of changes to my table data. As of now, there is an pgxn.org provided extension by which we could leverage this functionality, but when Temporal tables will become part of the main Postgresql release [and not as an extension] ? Even in the Postgresql 14 beta release notes it is not mentioned.
thanks,
Anand.
On Thursday, June 17, 2021, Anand Sowmithiran <anand.sow@gmail.com> wrote:
I am looking to use the temporal tables feature for keeping track of changes to my table data. As of now, there is an pgxn.org provided extension by which we could leverage this functionality, but when Temporal tables will become part of the main Postgresql release [and not as an extension] ? Even in the Postgresql 14 beta release notes it is not mentioned.
This email seems to provide a good starting point for reading up on the past work on the feature.
Its already too late for v14 since we are already in beta.
For a feature this complex its probably best to not have expectations of that nature.
And, since its not in the commitfest app, and the last emails I found are over a year old, it doesn’t seem to be a priority for anyone at this time.
David J.
On 6/17/21 11:10 PM, David G. Johnston wrote:
> And, since its not in the commitfest app, and the last emails I found
> are over a year old, it doesn’t seem to be a priority for anyone at this
> time.
I'm working on the application-time (aka valid-time) side of this, and
there is another patch for system-time. I can't speak to that patch, but
my patch is almost ready to share on the mailing list. I gave a progress
update at PGCon last month, but here is a summary of what it includes:
- You can define PERIODs (this is based heavily on Vik Fearing's patch
from a few years ago).
- You can define temporal PRIMARY KEYs and UNIQUE constraints.
- You can use FOR PORTION OF syntax to do temporal UPDATEs and DELETEs.
- You can define temporal FOREIGN KEYs including with all the
CASCADE/SET NULL/SET DEFAULT options.
I noticed a problem last week with the FOR PORTION OF code, so as soon
as I fix that I'm going to organize my changes into four patch files
(one for each of those bullet points) and send that to the list.
So far I haven't tried to support partitioned tables yet, so certainly
that will remain to be done. But that's the only missing piece I'm aware
of. (I'd be grateful for any help adding that btw. And Corey has already
offered to give me some help in general.)
I'm sorry it's been slow progress the last year. I do hope I can get
this landed for v15!
Yours,
--
Paul ~{:-)
pj@illuminatedcomputing.com
This is very helpful to know , we can likely count on temporal to be part of PG 15 !! I would be using Postgresql via Azure cloud[with Citus], so looking forward to use temporal features in my current development [using Extension] and then once it is part of PG 15, I guess that would make MSFT to start supporting pg15 with temporal on their cloud platform too.
thanks,
Anand.
On Fri, Jun 18, 2021 at 11:58 PM Paul Jungwirth <pj@illuminatedcomputing.com> wrote:
On 6/17/21 11:10 PM, David G. Johnston wrote:
> And, since its not in the commitfest app, and the last emails I found
> are over a year old, it doesn’t seem to be a priority for anyone at this
> time.
I'm working on the application-time (aka valid-time) side of this, and
there is another patch for system-time. I can't speak to that patch, but
my patch is almost ready to share on the mailing list. I gave a progress
update at PGCon last month, but here is a summary of what it includes:
- You can define PERIODs (this is based heavily on Vik Fearing's patch
from a few years ago).
- You can define temporal PRIMARY KEYs and UNIQUE constraints.
- You can use FOR PORTION OF syntax to do temporal UPDATEs and DELETEs.
- You can define temporal FOREIGN KEYs including with all the
CASCADE/SET NULL/SET DEFAULT options.
I noticed a problem last week with the FOR PORTION OF code, so as soon
as I fix that I'm going to organize my changes into four patch files
(one for each of those bullet points) and send that to the list.
So far I haven't tried to support partitioned tables yet, so certainly
that will remain to be done. But that's the only missing piece I'm aware
of. (I'd be grateful for any help adding that btw. And Corey has already
offered to give me some help in general.)
I'm sorry it's been slow progress the last year. I do hope I can get
this landed for v15!
Yours,
--
Paul ~{:-)
pj@illuminatedcomputing.com
Anand Sowmithiran <anand.sow@gmail.com> writes:
> This is very helpful to know , we can likely count on temporal to be part
> of PG 15 !!
You were just told *not* to count on that. Maybe it'll happen,
but I wouldn't take bets on it.
regards, tom lane
Sure, hence I wrote "likely"....!
On Sat, 19 Jun 2021, 8:10 pm Tom Lane, <tgl@sss.pgh.pa.us> wrote:
Anand Sowmithiran <anand.sow@gmail.com> writes:
> This is very helpful to know , we can likely count on temporal to be part
> of PG 15 !!
You were just told *not* to count on that. Maybe it'll happen,
but I wouldn't take bets on it.
regards, tom lane
"Hope" does not mean "likely".
On 6/19/21 11:24 AM, Anand Sowmithiran wrote:
Sure, hence I wrote "likely"....!On Sat, 19 Jun 2021, 8:10 pm Tom Lane, <tgl@sss.pgh.pa.us> wrote:Anand Sowmithiran <anand.sow@gmail.com> writes:
> This is very helpful to know , we can likely count on temporal to be part
> of PG 15 !!
You were just told *not* to count on that. Maybe it'll happen,
but I wouldn't take bets on it.
regards, tom lane
--
Angular momentum makes the world go 'round.
Angular momentum makes the world go 'round.