Fwd: Re: Add relcreated (timestamp) column to pg_class catalog to record the time an object was created - Mailing list pgsql-general

From melvin6925
Subject Fwd: Re: Add relcreated (timestamp) column to pg_class catalog to record the time an object was created
Date
Msg-id mbrynjrrq6ckoxap20n2gqx4.1461185428476@email.android.com
Whole thread Raw
Responses Re: Add relcreated (timestamp) column to pg_class catalog to record the time an object was created
List pgsql-general




Sent via the Samsung Galaxy S® 6, an AT&T 4G LTE smartphone
-------- Original message --------
From: "David G. Johnston" <david.g.johnston@gmail.com>
Date: 4/20/2016 16:13 (GMT-05:00)
To: Melvin Davidson <melvin6925@gmail.com>
Subject: Re: [GENERAL] Add relcreated (timestamp) column to pg_class catalog to record the time an object was created

try replying to the list....

On Wed, Apr 20, 2016 at 12:59 PM, Melvin Davidson <melvin6925@gmail.com> wrote:
To Tom:
"it *sounds* trivial, until you start thinking about backup/restore/replication situation"

That is BS. It is trivial.
In backup/restore, there is no need to record the create date, as a new create date is only needed if a table is dropped, otherwise if it exists on restore it is a data restore only situation. If it is to create a new database, then it is perfectly fine to use the new creation time. As for replication, then it is a straightforward duplication of create time.

To David:
"The burden seems to rest with you, not others.  I'll leave it at that since everything else has already been said elsewhere.
I'll add that failing to point out a previous discussion you were involved with is inconsiderate to others reading these lists. "

That is also a poor argument. My initial request was years ago and there has never been a valid argument as to why  this cannot be done. Please see my response to Tom Lane.
Further to my point:
http://stackoverflow.com/questions/4442323/how-to-find-out-when-a-particular-table-was-created-in-oracle
http://stackoverflow.com/questions/1171019/sql-server-table-creation-date-query
both show proof that this has been implemented in those databases.

So instead of replying back with "this has already been discussed and cannot be done", I respectfully request that I get the courtesy of a _detailed explanation_ of why this is appears so hard. It would also be nice if you allowed a couple of days for other users to comment before so rapidly dismissing it. Especially since we just went through a very length discussion on code of conduct.


On Wed, Apr 20, 2016 at 2:58 PM, David G. Johnston <david.g.johnston@gmail.com> wrote:
leaving off-list...

On Wednesday, April 20, 2016, melvin6925 <melvin6925@gmail.com> wrote:
You are correct, nothing has changed. This is a very simple request and should be easy to implement. I have yet to hear a valid, logical argument against it. In fact. This has been implemented in both Oracle and Sql Server. 

The burden seems to rest with you, not others.  I'll leave it at that since everything else has already been said elsewhere.

I'll add that failing to point out a previous discussion you were involved with is inconsiderate to others reading these lists. 

 I'll admit that maybe a better system for tracking and recording these kinds of requests would be nice - though expensive maintain - would be nice but in the meantime at least point to known history when bringing something like this up.  At worse it shows you did your research.

David J.




--
Melvin Davidson
I reserve the right to fantasize.  Whether or not you
wish to share my fantasy is entirely up to you.


pgsql-general by date:

Previous
From: rolf@winmutt.com
Date:
Subject: Re: Vacuum never completed....
Next
From: "David G. Johnston"
Date:
Subject: Re: Add relcreated (timestamp) column to pg_class catalog to record the time an object was created