<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 3.2//EN">
<HTML>
<HEAD>
<META HTTP-EQUIV="Content-Type" CONTENT="text/html; charset=iso-8859-1">
<META NAME="Generator" CONTENT="MS Exchange Server version 6.5.7638.1">
<TITLE>RE: [Mondrian] NULL literal</TITLE>
</HEAD>
<BODY>
<!-- Converted from text/plain format -->
<BR>
<P><FONT SIZE=2>On one of the web pages written by Mosha Pasumansky he<BR>
said that NULL/NULL == NULL, so I had assumed that<BR>
4/NULL and NULL/4 were also NULL. The reason I assumed that<BR>
was: if it were true, then running one of our MDX queries with and<BR>
without native crossjoin enabled returned the same result.<BR>
<BR>
I will backout the change.<BR>
<BR>
Richard<BR>
<BR>
-----Original Message-----<BR>
From: mondrian-bounces@pentaho.org on behalf of Julian Hyde<BR>
Sent: Wed 2/28/2007 5:55 PM<BR>
To: 'Mondrian developer mailing list'<BR>
Subject: RE: [Mondrian] NULL literal<BR>
<BR>
Richard,<BR>
<BR>
You only did half of what I suggested: you wrote testcases but you<BR>
didn't validate the results against MSAS.<BR>
<BR>
I ran<BR>
<BR>
with member [Measures].[Foo] as ' 4 / NULL '<BR>
select {[Measures].[Foo]} on 0<BR>
from [Sales]<BR>
<BR>
on MSAS 2K and it gave<BR>
<BR>
1.#INF<BR>
<BR>
as the answer. Now you've added a test which ensures the result is NULL,<BR>
i.e. wrong.<BR>
<BR>
Change 8802 is not an improvement. Please back it out.<BR>
<BR>
Julian<BR>
<BR>
<BR>
> -----Original Message-----<BR>
> From: mondrian-bounces@pentaho.org<BR>
> [<A HREF="mailto:mondrian-bounces@pentaho.org">mailto:mondrian-bounces@pentaho.org</A>] On Behalf Of Richard Emberson<BR>
> Sent: Wednesday, February 28, 2007 2:18 PM<BR>
> To: Mondrian developer mailing list<BR>
> Subject: Re: [Mondrian] NULL literal<BR>
><BR>
> In the 8802 checkin I added a NULL literal test method<BR>
> testNullLiteral() to BasicQueryTest.<BR>
><BR>
> Richard<BR>
><BR>
> Julian Hyde wrote:<BR>
> > <BR>
> ><BR>
> >> Interesting page on NULL and IS EMPTY<BR>
> >><BR>
> >><BR>
> ><BR>
> <A HREF="http://sqljunkies.com/WebLog/mosha/archive/2005/06/30/mdx_is_e">http://sqljunkies.com/WebLog/mosha/archive/2005/06/30/mdx_is_e</A><BR>
> mpty.aspx<BR>
> ><BR>
> > Looks like a page full of potential testcases to me!<BR>
> ><BR>
> > Anyone out there itching to contribute to mondrian by<BR>
> writing Q/A tests?<BR>
> > We could use your help!<BR>
> ><BR>
> > Julian<BR>
> ><BR>
> > _______________________________________________<BR>
> > Mondrian mailing list<BR>
> > Mondrian@pentaho.org<BR>
> > <A HREF="http://lists.pentaho.org/mailman/listinfo/mondrian">http://lists.pentaho.org/mailman/listinfo/mondrian</A><BR>
> ><BR>
><BR>
><BR>
> --<BR>
> Quis custodiet ipsos custodes:<BR>
> This email message is for the sole use of the intended<BR>
> recipient(s) and<BR>
> may contain confidential information. Any unauthorized review, use,<BR>
> disclosure or distribution is prohibited. If you are not the intended<BR>
> recipient, please contact the sender by reply email and destroy all<BR>
> copies of the original message.<BR>
> _______________________________________________<BR>
> Mondrian mailing list<BR>
> Mondrian@pentaho.org<BR>
> <A HREF="http://lists.pentaho.org/mailman/listinfo/mondrian">http://lists.pentaho.org/mailman/listinfo/mondrian</A><BR>
><BR>
<BR>
_______________________________________________<BR>
Mondrian mailing list<BR>
Mondrian@pentaho.org<BR>
<A HREF="http://lists.pentaho.org/mailman/listinfo/mondrian">http://lists.pentaho.org/mailman/listinfo/mondrian</A><BR>
<BR>
</FONT>
</P>
</BODY>
</HTML>