| From: | Simon Riggs <simon(at)2ndQuadrant(dot)com> |
|---|---|
| To: | Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us> |
| Cc: | Robert Haas <robertmhaas(at)gmail(dot)com>, Heikki Linnakangas <heikki(dot)linnakangas(at)enterprisedb(dot)com>, Andrew Dunstan <andrew(at)dunslane(dot)net>, Takahiro Itagaki <itagaki(dot)takahiro(at)oss(dot)ntt(dot)co(dot)jp>, Kjell Rune Skaaraas <kjella79(at)yahoo(dot)no>, pgsql-hackers(at)postgresql(dot)org |
| Subject: | Re: Add column if not exists (CINE) |
| Date: | 2010-07-22 00:43:01 |
| Message-ID: | 1279759381.1739.11169.camel@ebony |
| Views: | Raw Message | Whole Thread | Download mbox | Resend email |
| Thread: | |
| Lists: | pgsql-hackers |
On Wed, 2010-04-28 at 21:15 -0400, Tom Lane wrote:
> I still say
> that COR rather than CINE semantics would be appropriate for columns.
Viewed from a locking perspective, I would disagree.
COR semantics force a table rewrite, in certain cases. That makes it
hard to predict externally how long the command will run for.
As a DBA, I would want a command that executes without rewrite (if
appropriate) or does nothing.
Predictable behaviour is the most important concern.
That isn't necessarily an argument in favour of CINE, which seems
slightly less clear about what we might expect from that, in my reading
at least.
--
Simon Riggs www.2ndQuadrant.com
PostgreSQL Development, 24x7 Support, Training and Services
| From | Date | Subject | |
|---|---|---|---|
| Next Message | Robert Haas | 2010-07-22 00:48:59 | Re: Reworks of DML permission checks |
| Previous Message | KaiGai Kohei | 2010-07-22 00:13:30 | Re: ExecutorCheckPerms() hook |