Re: 7.0 bug ??

From: Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us>
To: claud yu <claud(at)tera(dot)cs(dot)tku(dot)edu(dot)tw>
Cc: pgsql-bugs(at)postgreSQL(dot)org
Subject: Re: 7.0 bug ??
Date: 2000-05-01 01:31:04
Message-ID: 24638.957144664@sss.pgh.pa.us
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-bugs

claud yu <claud(at)tera(dot)cs(dot)tku(dot)edu(dot)tw> writes:
> freechange=# insert into accountloginrec values (1,'test','test','test',NOW(),NOW()+30,Max(eshopcode)+1,null);
> ERROR: Attribute 'eshopcode' not found

> The same syntax at 6.5.3 is OK !!

Not a 7.0 bug, but a 6.5 bug --- it shouldn't have accepted that.
The correct SQL-approved way to do this sort of thing is

insert into accountloginrec
select 1,'test','test','test',NOW(),NOW()+30,Max(eshopcode)+1,null
from accountloginrec;

A values() expression is a standalone row-value constructor; it can't
legitimately contain any references to the insert destination table.
In fact the way INSERT is supposed to work is that it's processed as
INSERT INTO dest_table <something>
where the <something> is a row-value-yielding expression that would
have the same meaning whether it appeared in INSERT or not. 6.5 was
in error to accept free variables in the <something> as references
to the insert's destination table.

The INSERT...SELECT way works in 6.5 too, btw.

regards, tom lane

In response to

Browse pgsql-bugs by date

  From Date Subject
Next Message Rolf C Stadheim 2000-05-01 16:08:03 gmake[3]: *** No rule to make target `page/SUBSYS.o'. Stop. ??????????????????
Previous Message claud yu 2000-05-01 00:49:42 7.0 bug ??