Subject Re: [ib-support] Re: Ambiguous query or bug ?
Author Claudio Valderrama C.
""dianeb77x"" <dianeb77@...> wrote in message
>
> I hesitate (particularly in my unclear state) to ask, but "how is this
> opening a BCOW?"

Do you even hesitate??? <g>

My idea is not that I'm against what you posted, but that freedom and
flexibility has a limit where you allow the developer to shoot
himself/herself too easily, that's all.

It's easy to make a mistake by using the table instead of the alias when you
have room for ambiguity. In such context, it seems safer (albeit a bit
rigid) to enforce that if a table happens two times in a join, both
occurrences should be qualified by an alias. When one instance is the table
and other is the alias, it's preferable to have two alias so you know which
instance you are referring to.

Anyway, how does your answer fit with the issue that an alias should "hide"
the table name? Not in a join? There should be a couple of "not in this
case" in the standard, right?

Don't give up on the hard docos, Diane. The SQL doc I got from the
ansi-store answers only a few of my questions... and with an horrible
typeface. We need interpreters (like you) of the divine committee, envoys of
the gods, etc. (Jim's opinion on the committee is omitted for politeness.)

C.
--
Claudio Valderrama C. - http://www.cvalde.com - http://www.firebirdSql.org
Independent developer
Owner of the Interbase® WebRing