Re: [HACKERS] identity columns - Mailing list pgsql-hackers

From Michael Paquier
Subject Re: [HACKERS] identity columns
Date
Msg-id CAB7nPqTRuwadyG=f__KP5=Py-rEAYocWGVcECER0eLStG_1xOQ@mail.gmail.com
Whole thread Raw
In response to Re: [HACKERS] identity columns  (Vitaly Burovoy <vitaly.burovoy@gmail.com>)
List pgsql-hackers
On Thu, Apr 13, 2017 at 7:40 PM, Vitaly Burovoy
<vitaly.burovoy@gmail.com> wrote:
> On 4/6/17, Vitaly Burovoy <vitaly.burovoy@gmail.com> wrote:
>> On 4/6/17, Peter Eisentraut <peter.eisentraut@2ndquadrant.com> wrote:
>>> As I tried to mention earlier, it is very difficult to implement the IF
>>> NOT EXISTS behavior here, because we need to run the commands the create
>>> the sequence before we know whether we will need it.
>>
>> In fact with the function "get_attidentity" it is not so hard. Please,
>> find the patch attached.
> ...
>> I hope it is still possible to get rid of the "ADD GENERATED" syntax.
>
> Is it possible to commit the patch from the previous letter?
> It was sent before the feature freeze, introduces no new feature (only
> syntax change discussed in this thread and not implemented due to lack
> of a time of the author) and can be considered as a fix to the
> committed patch (similar to the second round in "sequence data type").

It seems to me that at least the part about removing ADD GENERATED
could be applied as an adjustment for the committed patch (INE would
be a new feature), and if there is a time to adjust already committed
features for the upcoming release, it is now. So I have added an open
item.
--
Michael



pgsql-hackers by date:

Previous
From: Michael Paquier
Date:
Subject: Re: [HACKERS] Quorum commit for multiple synchronous replication.
Next
From: Amit Langote
Date:
Subject: Re: [HACKERS] Declarative partitioning vs. information_schema