Re: [HACKERS] sequence creation\ - Mailing list pgsql-hackers

From darcy@druid.net (D'Arcy J.M. Cain)
Subject Re: [HACKERS] sequence creation\
Date
Msg-id m0z8yfa-00006FC@druid.net
Whole thread Raw
In response to Re: [HACKERS] sequence creation  ("Thomas G. Lockhart" <lockhart@alumni.caltech.edu>)
List pgsql-hackers
Thus spake Thomas G. Lockhart
> > Perhaps use the data type SERIAL to autocreate a sequence.  Should
> > make life easier for novices.  We are getting too many sequence
> > questions.
>
> That would be possible. I'd be happier doing it for v6.5, since I'm
> hoping to work on docs in the meantime. Of course, maybe it would be
> easy :)
>
> What should the syntax be exactly?
>
>   CREATE TABLE t (s SERIAL);
>
> or
>
>   CREATE TABLE t (i INT DEFAULT SERIAL);

I think the first one is simpler and, as far as I know, is more common
in existing practice.

> Are there alternate syntaxes from other DBs which should be considered?

The standard doesn't specify a serial type.  One question would be,
what is the behaviour?  Should the field always be set to the next
number on insert or should we be able to override it?  I prefer the
former but I can see that it would be a problem on dump and reload.
How do other DBs handle this?

--
D'Arcy J.M. Cain <darcy@{druid|vex}.net>   |  Democracy is three wolves
http://www.druid.net/darcy/                |  and a sheep voting on
+1 416 424 2871     (DoD#0082)    (eNTP)   |  what's for dinner.

pgsql-hackers by date:

Previous
From: "Thomas G. Lockhart"
Date:
Subject: Re: [HACKERS] sequence creation
Next
From: Bruce Momjian
Date:
Subject: system caches