Re: Domain check taking place unnecessarily? - Mailing list pgsql-performance

From Mark Hills
Subject Re: Domain check taking place unnecessarily?
Date
Msg-id 2302091056380.14752@stax.localdomain
Whole thread Raw
In response to Re: Domain check taking place unnecessarily?  ("David G. Johnston" <david.g.johnston@gmail.com>)
List pgsql-performance
On Wed, 8 Feb 2023, David G. Johnston wrote:

> On Wed, Feb 8, 2023 at 11:01 AM Mark Hills <mark@xwax.org> wrote:
> 
> >
> > CREATE DOMAIN hash AS text
> >     CHECK (VALUE ~ E'^[a-zA-Z0-9]{8,32}$');
> >
> > devstats=> ALTER TABLE invite ADD COLUMN test hash;
> > ALTER TABLE
> > Time: 30923.380 ms (00:30.923)
> >
> 
> Necessarily, I presume because if you decided that the check on the domain
> should be "value is not null" (don't do this though...) the column addition
> would have to fail for existing rows (modulo defaults...).

I'm not sure I'm parsing this paragraph correctly, but the existing rows 
don't provide any data to the domain check. Perhaps you could clarify.

Many thanks

-- 
Mark



pgsql-performance by date:

Previous
From: Mark Hills
Date:
Subject: Re: Domain check taking place unnecessarily?
Next
From: Tom Lane
Date:
Subject: Re: Domain check taking place unnecessarily?