Date:         Wed, 19 Nov 2003 09:36:02 +1100
Reply-To:     Simon Freidin <s.freidin@iaesr.unimelb.edu.au>
Sender:       "SPSSX(r) Discussion" <SPSSX-L@LISTSERV.UGA.EDU>
From:         Simon Freidin <s.freidin@iaesr.unimelb.edu.au>
Subject:      Re: Write variable name into new variable
Comments: To: "Steiner, Markus" <markus.steiner@uba.de>
In-Reply-To:  <7677B11D9B2F8E429EE90D36C2228F0602985E8A@hermes-rz8.uba.de >
Content-Type: text/plain; charset="us-ascii"; format=flowed

data list list/ID (f1) loc  (a1) date (a6)  x y z (3f3).
begin data.
1   N   12.11.  12     22     53
2   W   23.10.  34     11     34
3   N   24.08.  14     31     76
end data.

varstocases /make value from x y z/index=name(value).
list.

Output:
ID LOC DATE   NAME     VALUE

  1 N   12.11. X          12
  1 N   12.11. Y          22
  1 N   12.11. Z          53
  2 W   23.10. X          34
  2 W   23.10. Y          11
  2 W   23.10. Z          34
  3 N   24.08. X          14
  3 N   24.08. Y          31
  3 N   24.08. Z          76

Unfortunately you have to choose between getting COL or NAME created by the
index subcommand, it won't do both simultaneously.
Anyone have a workaround?

To create COL:
varstocases /make value from x y z/index=col.

cheers
Simon

At 04:17 PM 11/18/2003 +0100, Steiner, Markus wrote:
>Hello,
>
>while partly transposing a data set I would need to write the variable
>name of some variables into a new variable, so that the information, from
>which variable are the values after transposition are still available.
>
>Here an example:
>
>Original data
>
>ID loc  date    x      y      z
>1   N   12.11.  12     22     53
>2   W   23.10.  34     11     34
>3   N   24.08.  14     31     76
>
>with the folowing code:
>
>
>DATA LIST FREE / id (N) loc (A) date (EDATE) var1 (N) var2 (N) var3 (N).
>BEGIN DATA
>1   N   12.11.03  12     22     53
>2   W   23.10.03  34     11     34
>3   N   24.08.03  14     31     76
>END DATA.
>
>IF ( MISSING(LAG($CASENUM)) OR id <> LAG(id) ) ROW=1.
>IF(MISSING(ROW)) ROW=LAG(ROW)+1 .
>VECTOR VAR=var1 TO var3 .
>LOOP COL=1 TO 3 .
>COMPUTE value=VAR(COL).
>XSAVE OUTFILE 'tmp_flip.sav' / KEEP id loc date row col value .
>END LOOP.
>EXECUTE.
>GET FILE 'tmp_flip.sav'.
>
>it looks like:
>
>ID loc  date   row  col  value
>1   N   12.11.  1    1    12
>1   N   12.11.  1    2    22
>1   N   12.11.  1    3    53
>2   W   23.10.  1    1    34
>2   W   23.10.  1    2    11
>2   W   23.10.  1    3    34
>3   N   24.08.  1    1    14
>3   N   24.08.  1    2    31
>3   N   24.08.  1    3    76
>
>
>and that's what I want:
>
>ID loc  date   name col  value
>1   N   12.11.  X    1    12
>1   N   12.11.  Y    2    22
>1   N   12.11.  Z    3    53
>2   W   23.10.  X    1    34
>2   W   23.10.  Y    2    11
>2   W   23.10.  Z    3    34
>3   N   24.08.  X    1    14
>3   N   24.08.  Y    2    31
>3   N   24.08.  Z    3    76
>
>Any help would be very helpful.
>
>Regards.
>
>Markus


Research Database Manager and Analyst
Melbourne Institute of Applied Economic and Social Research
The University of Melbourne
Melbourne VIC 3010 Australia
Tel: (03) 8344 3601 Fax: (03) 8344 5630
http://www.melbourneinstitute.com/hilda/