Re: SQL Update with a Join Syntax

From: Mark D Powell <Mark.Powell_at_eds.com>
Date: 13 Aug 2004 14:11:38 -0700
Message-ID: <2687bb95.0408131311.3b228b68_at_posting.google.com>


jcelko212_at_earthlink.net (--CELKO--) wrote in message news:<18c7b3c2.0408121154.3b0949a6_at_posting.google.com>...
> >> What do you mean that no correlation name is allowed? You can
> place a table name (correlation) label on an update, otherwise how
> would you perform co-ordingated subqueries? (Code runs on Oracle 7.0 -
> 9.2) <<
>
> NOT in Standard SQL; look it up!
>
> "Caesar: Pardon him, Theodotus. He is a barbarian and thinks the
> customs of his tribe and island are the laws of nature." - Caesar and
> Cleopatra; George Bernard Shaw 1898
>
> It would make no sense. The ANSI/ISO model is that a correlation
> creates a temporaty working table with that name and the data of the
> table expression on the left hand side of the AS operator. You also
> have the option of renaming the columns in the new table.
>
> This means that the working table, not the base table, would be
> updated and then disappear.

This is probably just a terminology problem, more likely on my part, on what is meant by correlation. I looked in the MS SQL manual and they also support coordinated subquery on an update. I am sure this also works in DB2 so it seems unlikely that the table label "A" in my example represents any type of correlation as meant by the standard.

  • Mark D Powell --
Received on Fri Aug 13 2004 - 23:11:38 CEST

Original text of this message