Path: utzoo!mnetor!uunet!seismo!sundc!pitstop!sun!decwrl!pyramid!voder!blia!blipyramid!mike
From: mike@blipyramid.BLI.COM (Mike Ubell)
Newsgroups: comp.databases
Subject: Re: Informix 4GL Question?
Message-ID: <62@blipyramid.BLI.COM>
Date: 25 Feb 88 14:45:08 GMT
References: <714@uel.uel.co.uk> <2314@geac.UUCP> <974@pasteur.Berkeley.Edu>
Organization: Britton Lee, Inc.
Lines: 19
Keywords: Informix 4GL
Summary: aggregate is not correct

In article <974@pasteur.Berkeley.Edu>, larry@postgres.uucp (Larry Rowe) writes:
> the referenced messages asked about relational calculus queries that
> answer the queries:
> 	1. get the first occurrence of some predicate
> 	2. get the first N occurrences of some predicate
> 	3. return true/false if an occurrence exists
> QUEL has an aggregate function that does numbers 1 and 3.  the ANY
> this illustrates query 3.  query 1 is simple to do, just put the ANY
> aggregate in the predicate rather than the target list.  for example,
> 
> 	RETRIEVE (emp.all)
> 	WHERE ANY(emp.name by emp.dept where emp.dept= "toy") = 1
> 
> the by-clause is needed to link the tuple variable in the ANY
> aggregate to the record constructed in the target list.

The above aggregate returns all emp records where emp.dept = "toy".
Quel joins the by list to ALL records in the original relation not
just the first one.