Date:         Thu, 3 Sep 1998 14:09:07 -0500
Reply-To:     etverdek@spss.com
Sender:       "SPSSX(r) Discussion" <SPSSX-L@UGA.CC.UGA.EDU>
From:         Edward Tverdek <etverdek@SPSS.COM>
Organization: SPSS Inc.
Subject:      Re: reorganizing data sets & vectors/arrays
Content-Type: text/plain; charset=us-ascii

I responded to this yesterday but still don't see it posted, so I'm
responding again.  Forgive me the the first makes it through and this
is redundant.
 
I asked our webmaster to post to the SPSS Base FAQ sample jobs which
illustrate both of the procedures that Frank describes.  They're
already up; you can find them under the heading "Restructuring the
case base of an SPSS file" at
 
http://www.spss.com/tech/FAQ-SPSS.html#Q3
 
I hope this helps,
 
Ed Tverdek
SPSS Technical Support
 
 
Frank H. Jurden wrote:
>
> several listers have posted questions regarding the problem of
> transposing data *by case* in spss.  for the uninitiated, this problem
> is one of transforming the "traditional data set" (consisting of a
> single record for each respondent and multple columns for the variables)
> into one in which the respondents are "stacked" (viz., the same
> respondent may occupy multiple rows).  so if the original data set
> contains N cases and p variables, the transposed data set will contain N
> x p rows (and some number of columns).
>
> on occassion, the opposite is required (taking multiple rows for each
> respondent and "stringing them out" across multiple columns.
>
> althought spss does have a transpose function, it can not be done by
> case (btw, sas does).
>
> i have faced this problem several times, and it took quite a lot of
> digging and debugging to get the code to work (sorry folks - no menu
> options here!).  it involves a combination of the "input program" and
> "loop - end loop" syntax (copies of which i can post if anyone wants).
>
> i'd like to see spss incorporate this into their FAQ or on their
> website.