Re: Shouldn't this be an error? - Mailing list pgsql-hackers

From Bruce Momjian
Subject Re: Shouldn't this be an error?
Date
Msg-id 200101291658.LAA17063@candle.pha.pa.us
Whole thread Raw
In response to Re: Shouldn't this be an error?  (Tom Lane <tgl@sss.pgh.pa.us>)
List pgsql-hackers
> Vince Vielhaber <vev@michvhf.com> writes:
> > Shouldn't this insert fail?   This is 7.0.3.
> 
> > template1=# create table foo(x int,y varchar(10),z datetime);
> > CREATE
> > template1=# insert into foo(x,y,z) values(1,'asdf');
> > INSERT 19222 1
> 
> We've always allowed trailing columns to be omitted, whether a column
> name list is specified or not.  This is not per spec --- SQL92 and SQL99
> both say that all the columns must be provided --- but I'm rather
> hesitant to enforce the spec's stricter rule at this point.  Seems like
> it'd probably break some existing apps.
> 
> A compromise position would be to allow dropping trailing columns only
> when the column name list is omitted.

Yikes, we allow the column to be missing even if they specify it in
INSERT.  That seems very wrong.

--  Bruce Momjian                        |  http://candle.pha.pa.us pgman@candle.pha.pa.us               |  (610)
853-3000+  If your life is a hard drive,     |  830 Blythe Avenue +  Christ can be your backup.        |  Drexel Hill,
Pennsylvania19026
 


pgsql-hackers by date:

Previous
From: "Ross J. Reedstrom"
Date:
Subject: Re: Sure enough, the lock file is gone
Next
From: Zeugswetter Andreas SB
Date:
Subject: AW: Shouldn't this be an error?