From: | Bruce Momjian <pgman(at)candle(dot)pha(dot)pa(dot)us> |
---|---|
To: | Bradley Kieser <brad(at)kieser(dot)net> |
Cc: | billy(at)oconnoronline(dot)net, DCorbit(at)connx(dot)com, pgsql-hackers(at)postgresql(dot)org |
Subject: | Re: Feature request: Truncate table |
Date: | 2002-06-12 22:26:18 |
Message-ID: | 200206122226.g5CMQIE18056@candle.pha.pa.us |
Views: | Raw Message | Whole Thread | Download mbox | Resend email |
Thread: | |
Lists: | pgsql-hackers |
Bradley Kieser wrote:
> Well in Ingres there is a WORLD of difference! For a start, you don't
> lock out the system catalog. Secondly it is an unlogged event, so it
> beats "delete from table_name" hands down! Then, of course, it preserves
> all permissions, you keep the same OID, so views, et al, can remain in
> tact, as with other objects that referece it.
>
> These are very important considerations in real-world applications esp.
> when a large number of objects may reference the table.
>
>
> Which brings me to another point - I would dearly love to see a
> "refresh" option based on object name added to the system. This would
> check all references to a dropped object, by name, and repoint them to
> the new instance of that object (i.e. if you do a drop/create, it
> doesn't mess up your entire system if you forgot about a view or three!).
We have actually be moving away from name-based linking so you can
rename tables and things still work. I can see value in a relinking
system, but we would have to know the old oid and new name, I guess.
--
Bruce Momjian | http://candle.pha.pa.us
pgman(at)candle(dot)pha(dot)pa(dot)us | (610) 853-3000
+ If your life is a hard drive, | 830 Blythe Avenue
+ Christ can be your backup. | Drexel Hill, Pennsylvania 19026
From | Date | Subject | |
---|---|---|---|
Next Message | Bruce Momjian | 2002-06-12 22:30:06 | Re: Integrating libpqxx |
Previous Message | Jeroen T. Vermeulen | 2002-06-12 22:25:41 | Re: Integrating libpqxx |