Re: Date created for tables

On 12/24/19 10:39 AM, Adrian Klaver wrote:
> On 12/23/19 6:14 PM, Ron wrote:
>> On 12/23/19 7:01 PM, Bruce Momjian wrote:
>>> On Thu, Dec  5, 2019 at 05:10:20PM +0000, Chloe Dives wrote:
>>>> Having moved to PostgreSQL from Oracle a few years ago I have been
>>>> generally
>>>> very impressed by Postgres, but there are a few things that I still
>>>> miss. One
>>>> of those is being able to see the created and last modified dates for
>>>> database
>>>> objects.
>>>>
>>>>
>>>> Is this something that has been considered for implementation?
>>> I wrote a blog about this:
>>>
>>>     https://momjian.us/main/blogs/pgblog/2017.html#November_21_2017
>>
>> You all are *grossly* over-complicating this.
>
> Not really. This discussion has come up before and it starts with the
> simple case of timestamp the initial CREATE. This would suffice for some
> folks. However, it then progresses into a request for full object audit
> system.

This is directly akin to Henry Ford refusing to build cars because people
will *someday*** want computerized fuel injection, crumple zones and air bags.

> I understand why there is no great desire to start down this path by the
> developers, they know the pressure would be on to expand the code. As
> Fabrízio mentions in another post this is something that could be covered
> in an extension. FYI, I do it by using Sqitch for my schema object creation.
>
>>
>> By creation time, "we DBAs" think the time we ran "CREATE object", not
>> when pg_dump, pg_basebackup and pg_update ran.
>>
>> Likewise, modification time is when we last ran an ALTER command ran, not
>> when VACUUM ran (that's tracked elsewhere) or DML ran.
>>
>> That's all.
>>
>> --
>> Angular momentum makes the world go 'round.
>
>

--
Angular momentum makes the world go 'round.