Re: ALTER TEXT field to VARCHAR(1024) - Mailing list pgsql-general

From Merlin Moncure
Subject Re: ALTER TEXT field to VARCHAR(1024)
Date
Msg-id CAHyXU0ywZwgbkB213fB1P1nViPsB7vq70HoZPWf99pj8-35PoQ@mail.gmail.com
Whole thread Raw
In response to Re: ALTER TEXT field to VARCHAR(1024)  (Bill Moran <wmoran@potentialtech.com>)
Responses Re: ALTER TEXT field to VARCHAR(1024)  (John McKown <john.archie.mckown@gmail.com>)
List pgsql-general
On Fri, Sep 19, 2014 at 7:16 AM, Bill Moran <wmoran@potentialtech.com> wrote:
> On Fri, 19 Sep 2014 09:32:09 +0200
> Marius Grama <mariusneo@gmail.com> wrote:
>> Can anybody explain me what happens in the background when the alter
>> statement is executed? I've tried it out on a small copy of the table (70K)
>> and the operation completed in 0.2 seconds.
>> Will the table be completely locked during the execution of the ALTER
>> statement?
>
> I share Gavin's concern that you're fixing this in the wrong place.  I expect
> that you'll be better served by configuring the middleware to do the right thing.

I'll pile on here: in almost 20 years of professional database
development I've never had an actual problem that was solved by
introducing or shortening a length constraint to text columns except
in cases where overlong strings violate the data model (like a two
character state code for example).  It's a database equivalent of "C
programmer's disease".  Input checks from untrusted actors should
happen in the application.

merlin


pgsql-general by date:

Previous
From: Merlin Moncure
Date:
Subject: Re: Postgre SQL SHA-256 Compliance
Next
From: Stephen Frost
Date:
Subject: Re: Postgre SQL SHA-256 Compliance