Path: utzoo!utgpu!jarvis.csri.toronto.edu!rutgers!mailrus!tut.cis.ohio-state.edu!cs.utexas.edu!uunet!cuc1!lcain
From: lcain@cuc1.UUCP (Leroy Cain)
Newsgroups: comp.databases
Subject: Re: Sybase bug?
Summary: Other RDBMS work
Message-ID: <2098@cuc1.UUCP>
Date: 5 Jul 89 05:10:52 GMT
References: <613@kylie.oz> <571@daitc.daitc.mil>
Organization: Columbia Union College; Takoma Park, MD 20912
Lines: 44

In article <571@daitc.daitc.mil>, jkrueger@daitc.daitc.mil (Jonathan Krueger) writes:
> In article <613@kylie.oz>, christie@kylie (Chris Tham) writes:
> >I found the following anomaly and wondered if it is a bug in Sybase or
> >normal in SQL:  [test case looks like:
> 
> 	+-------+			+-------+
> 	| hello |    a          	| world |    b
> 	+-------+------+		+-------+------+
> 		|    1 |			+------+
> 		+------+
> 
> 	select * from hello where a = 1 or a in (select b from world);
> 
> 			+--------+
> 			| result |    a
> 			+--------+------+
> 				 +------+
> 
> It's normal.  The semantics are usually unintended, however.  The
> subquery implies a join to an empty table.  This means a cartesian
> product where one matrix has size zero.  Or if you prefer, in the
> relational model intersection is undefined on two relations if either
> is empty.


That is an interesting explanation but I don't buy it.  I checked
Informix-SQL and Unify 3.2 and they both product the correct results.
 
 			+--------+
 			| result |    a
 			+--------+------+
				 |   1  |
 				 +------+

[ example of union delete ]

So why the example using a union ( which is correct ) if you think there is
an intersection?  Or am I just misunderstand you?


MS-DOS Just say NO!!!!!			      OS/2 Why????
Leroy Cain;      Columbia Union College;      Mathematical Sciences Department
7600 Flower Ave. WH406;  	              Takoma Park, Md 20912
(301) 891-4172				      uunet!cucstud!lcain