Re: PSPP-BUG: Cant compile latest stable version pspp-stable-200910XX

[Leonard Evens]

On Thu, 2009-10-29 at 10:21 -0400, Jason Stover wrote:
> On Thu, Oct 29, 2009 at 08:48:37AM -0500, Leonard Evens wrote:
> > Okay.  I compiled 0.6.2.  But it has the same crosstab bug I reported
> > before for 0.6.1.  If I do a crosstab on split data, then in the second
> > part of the split, it reports the total number of missing entries rather
> > than the number missing just for that part of the split. (Possibly a
> > variable was not reinitialized?)  I had hoped that I could find a
> > version where that bug had been fixed.
> 
> Can you send a sample data file that shows the problem?

Try the attached .sav file.  Split on the values of the variable ven3,
which separates data with ven = 3 from data with ven not= 3.  Then do a
crosstab on any pair of variables, e.g,
gender vs zipcode.  One of  the two results will report the wrong number
of missing entries and the wrong number of total entries.  Which one is
wrong will depend on the ordering of the data.   The total should be 178
for venue 3 and 68 for all other venues combined.  The total number of
entries independent of venue is 246.

I also attached a file showing the ou tpu t I got for age vs zipcode,
and I indicated in that file where the numbers reported were wrong.
-- 
Leonard Evens address@hidden
Professor Emeritus, Department of Mathematics, Northwestern University

NSVSurveyNumbersData_Ven3.sav
Description: application/extension-sav

cross_tab_err.txt
Description: Text document