SPSSX Discussion

Adding missing cases to univariate data file

classic Classic list List threaded Threaded
♦
♦
 5 messages Options
Jake Gross
Reply | Threaded
Open this post in threaded view
♦
♦
|

Adding missing cases to univariate data file

Jake Gross
Hi all,

Another question for you.

I am working with longitudinal annual data of the type below.

DATA LIST LIST /id year var1 var2.
BEGIN DATA
1 2000 2 3
1 2001 5 2
1 2002 8 2
2 2000 2 5
2 2002 3 3
2 2003 8 5
3 2004 9 6
4 2005 0 2
END DATA.
LIST.

I would like to restructure the file so that I create a new case for all
years from 1999-2005...something like below.

DATA LIST LIST /id year var1 var2.
BEGIN DATA
1 1999
1 2000 2 3
1 2001 5 2
1 2002 8 2
1 2003
1 2004
1 2005
2 1999
2 2000 2 5
2 2001
2 2002 3 3
2 2003 8 5
2 2004
2 2005
3 1999
3 2000
3 2001
3 2002
3 2004 9 6
3 2005
4 1999
4 2000
4 2001
4 2002
4 2003
4 2004
4 2005 0 2
END DATA.
LIST.

Any suggestions about how I might tackle this?

Peace,
Jake

=====================
To manage your subscription to SPSSX-L, send a message to
[hidden email] (not to SPSSX-L), with no body text except the
command. To leave the list, send the command
SIGNOFF SPSSX-L
For a list of commands to manage subscriptions, send the command
INFO REFCARD
Richard Ristow
Reply | Threaded
Open this post in threaded view
♦
♦
|

Re: Adding missing cases to univariate data file

Richard Ristow
At 01:31 PM 4/2/2008, Jake Gross wrote:

>I am working with longitudinal annual data of the type below.
|-----------------------------|---------------------------|
|Output Created               |02-APR-2008 15:56:39       |
|-----------------------------|---------------------------|
[TestInpt]

id year var1 var2

  1 2000    2    3
  1 2001    5    2
  1 2002    8    2
  2 2000    2    5
  2 2002    3    3
  2 2003    8    5
  3 2004    9    6
  4 2005    0    2

Number of cases read:  8    Number of cases listed:  8

>I would like to restructure the file so that I create a new case for all
>years from 1999-2005, [with missing values for data in the added cases]

Mildly tricky, since you can't use INPUT PROGRAM/LOOP/END CASE when
reading an existing SPSS file (sigh). You could use LOOP/XSAVE to get
an effect similar to END CASE's, and I'd do that if XSAVE could write
to a dataset (second sigh).

Here's a VARSTOCASES solution (code and output not saved separately):

DATASET DECLARE   YearList.

AGGREGATE OUTFILE=YearList
   /BREAK=ID
   /NRECS 'No. of records for ID, but don''t really care' = NU.

DATASET ACTIVATE  YearList WINDOW=FRONT.
.  /**/ LIST /*-*/.

List
|-----------------------------|---------------------------|
|Output Created               |02-APR-2008 16:13:33       |
|-----------------------------|---------------------------|
[YearList]

id   NRECS

  1       3
  2       3
  3       1
  4       1

Number of cases read:  4    Number of cases listed:  4


NUMERIC    YEAR1999 TO YEAR2005 (F4).
DO REPEAT  VARIABLE = YEAR1999 TO YEAR2005
           /VALUE    =     1999 TO     2005.
.  COMPUTE VARIABLE = VALUE.
END REPEAT.

.  /**/ LIST /*-*/.

List
|-----------------------------|---------------------------|
|Output Created               |02-APR-2008 16:13:33       |
|-----------------------------|---------------------------|
[YearList]

id   NRECS YEAR1999 YEAR2000 YEAR2001 YEAR2002 YEAR2003 YEAR2004 YEAR2005

  1       3   1999     2000     2001     2002     2003     2004     2005
  2       3   1999     2000     2001     2002     2003     2004     2005
  3       1   1999     2000     2001     2002     2003     2004     2005
  4       1   1999     2000     2001     2002     2003     2004     2005

Number of cases read:  4    Number of cases listed:  4


VARSTOCASES
   /MAKE YEAR FROM YEAR1999 YEAR2000 YEAR2001 YEAR2002 YEAR2003
                   YEAR2004 YEAR2005
  /KEEP =  id
  /NULL = KEEP.

Variables to Cases
|-----------------------------|---------------------------|
|Output Created               |02-APR-2008 16:13:33       |
|-----------------------------|---------------------------|
[YearList]

Generated Variables
|----|------|
|Name|Label |
|----|------|
|YEAR|<none>|
|----|------|

Processing Statistics
|-------------|-|
|Variables In |9|
|Variables Out|2|
|-------------|-|


.  /**/ LIST /*-*/.
List
|-----------------------------|---------------------------|
|Output Created               |02-APR-2008 16:13:33       |
|-----------------------------|---------------------------|
[YearList]
  id YEAR
[List of cases suppressed]
Number of cases read:  28    Number of cases listed:  28


MATCH FILES
    /FILE=YearList
    /FILE=TestInpt
    /BY ID YEAR.
LIST.

List
|-----------------------------|---------------------------|
|Output Created               |02-APR-2008 16:13:34       |
|-----------------------------|---------------------------|
id YEAR var1 var2

  1 1999    .    .
  1 2000    2    3
  1 2001    5    2
  1 2002    8    2
  1 2003    .    .
  1 2004    .    .
  1 2005    .    .
  2 1999    .    .
  2 2000    2    5
  2 2001    .    .
  2 2002    3    3
  2 2003    8    5
  2 2004    .    .
  2 2005    .    .
  3 1999    .    .
  3 2000    .    .
  3 2001    .    .
  3 2002    .    .
  3 2003    .    .
  3 2004    9    6
  3 2005    .    .
  4 1999    .    .
  4 2000    .    .
  4 2001    .    .
  4 2002    .    .
  4 2003    .    .
  4 2004    .    .
  4 2005    0    2

Number of cases read:  28    Number of cases listed:  28
=================================
APPENDIX: Test data, and all code
(Test data from original posting,
and very nicely done)
=================================
DATA LIST LIST /id year var1 var2.
BEGIN DATA
1 2000 2 3
1 2001 5 2
1 2002 8 2
2 2000 2 5
2 2002 3 3
2 2003 8 5
3 2004 9 6
4 2005 0 2
END DATA.
FORMATS id         (F2)
         year       (F4)
         var1  var2 (F3).
DATASET NAME      TestInpt  WINDOW=FRONT.
LIST.

DATASET DECLARE   YearList.
AGGREGATE OUTFILE=YearList
   /BREAK=ID
   /NRECS 'No. of records for ID, but don''t really care' = NU.
DATASET ACTIVATE  YearList WINDOW=FRONT.

.  /**/ LIST /*-*/.

NUMERIC    YEAR1999 TO YEAR2005 (F4).
DO REPEAT  VARIABLE = YEAR1999 TO YEAR2005
           /VALUE    =     1999 TO     2005.
.  COMPUTE VARIABLE = VALUE.
END REPEAT.

.  /**/ LIST /*-*/.

VARSTOCASES
   /MAKE YEAR FROM YEAR1999 YEAR2000 YEAR2001 YEAR2002 YEAR2003
                   YEAR2004 YEAR2005
  /KEEP =  id
  /NULL = KEEP.

.  /**/ LIST /*-*/.

MATCH FILES
    /FILE=YearList
    /FILE=TestInpt
    /BY ID YEAR.
LIST.

=====================
To manage your subscription to SPSSX-L, send a message to
[hidden email] (not to SPSSX-L), with no body text except the
command. To leave the list, send the command
SIGNOFF SPSSX-L
For a list of commands to manage subscriptions, send the command
INFO REFCARD
Jake Gross
Reply | Threaded
Open this post in threaded view
♦
♦
|

Re: Adding missing cases to univariate data file

Jake Gross
Richard & SPSS Folks,

Thanks. This worked like a charm. I also found some code that I adapted (see below) to do the trick. I have another puzzle for you and others if anyone is willing.

For the same data set I would like to insert the missing cases from the first year of data forward to 2005. So, for example with ID#1 below I would want to insert 2003, 2004 and 2005 whereas with ID#3 I would only want to insert 2005. Any thoughts building from the code I used or from the code you sent? I am thinking a macro might work or something that would set the first value in loop equal to the first year in the dataset, but haven't hit on a way to do it.

DATA LIST LIST/ caseid YearOnly var1 var2.
BEGIN DATA
1 2000 2 3
1 2001 5 2
1 2002 8 2
2 2000 2 5
2 2002 3 3
2 2003 8 5
3 2004 9 6
4 2005 0 2.
END DATA.

SORT CASES by CASEID (A) YEARONLY (A).
SAVE OUTFILE= 'C:\temp.sav'.
SORT CASES by CASEID Year.
AGGREGATE
  /OUTFILE=*
  /BREAK=CASEID
  /YearOnly = NU(YearOnly).
LOOP yr = 1999 TO 2005.
 + XSAVE OUTFILE='temp.sav'
 /KEEP=CASEID yr .
END LOOP.
EXECUTE.
GET FILE='temp.sav'.
RENAME VARIABLE (yr=YearOnly).
****Matching the Files and filling in the gaps in years.
MATCH FILES /FILE=*
  /FILE='C:\temp.sav'
  /BY CASEID YearOnly.
EXECUTE.
SAVE OUTFILE= 'C:\temp2.sav'.

________________________________________
From: Richard Ristow [[hidden email]]
Sent: Wednesday, April 02, 2008 3:19 PM
To: Gross, Paul Jacob; [hidden email]
Subject: Re: Adding missing cases to univariate data file

At 01:31 PM 4/2/2008, Jake Gross wrote:

>I am working with longitudinal annual data of the type below.
|-----------------------------|---------------------------|
|Output Created               |02-APR-2008 15:56:39       |
|-----------------------------|---------------------------|
[TestInpt]

id year var1 var2

  1 2000    2    3
  1 2001    5    2
  1 2002    8    2
  2 2000    2    5
  2 2002    3    3
  2 2003    8    5
  3 2004    9    6
  4 2005    0    2

Number of cases read:  8    Number of cases listed:  8

>I would like to restructure the file so that I create a new case for all
>years from 1999-2005, [with missing values for data in the added cases]

Mildly tricky, since you can't use INPUT PROGRAM/LOOP/END CASE when
reading an existing SPSS file (sigh). You could use LOOP/XSAVE to get
an effect similar to END CASE's, and I'd do that if XSAVE could write
to a dataset (second sigh).

Here's a VARSTOCASES solution (code and output not saved separately):

DATASET DECLARE   YearList.

AGGREGATE OUTFILE=YearList
   /BREAK=ID
   /NRECS 'No. of records for ID, but don''t really care' = NU.

DATASET ACTIVATE  YearList WINDOW=FRONT.
.  /**/ LIST /*-*/.

List
|-----------------------------|---------------------------|
|Output Created               |02-APR-2008 16:13:33       |
|-----------------------------|---------------------------|
[YearList]

id   NRECS

  1       3
  2       3
  3       1
  4       1

Number of cases read:  4    Number of cases listed:  4


NUMERIC    YEAR1999 TO YEAR2005 (F4).
DO REPEAT  VARIABLE = YEAR1999 TO YEAR2005
           /VALUE    =     1999 TO     2005.
.  COMPUTE VARIABLE = VALUE.
END REPEAT.

.  /**/ LIST /*-*/.

List
|-----------------------------|---------------------------|
|Output Created               |02-APR-2008 16:13:33       |
|-----------------------------|---------------------------|
[YearList]

id   NRECS YEAR1999 YEAR2000 YEAR2001 YEAR2002 YEAR2003 YEAR2004 YEAR2005

  1       3   1999     2000     2001     2002     2003     2004     2005
  2       3   1999     2000     2001     2002     2003     2004     2005
  3       1   1999     2000     2001     2002     2003     2004     2005
  4       1   1999     2000     2001     2002     2003     2004     2005

Number of cases read:  4    Number of cases listed:  4


VARSTOCASES
   /MAKE YEAR FROM YEAR1999 YEAR2000 YEAR2001 YEAR2002 YEAR2003
                   YEAR2004 YEAR2005
  /KEEP =  id
  /NULL = KEEP.

Variables to Cases
|-----------------------------|---------------------------|
|Output Created               |02-APR-2008 16:13:33       |
|-----------------------------|---------------------------|
[YearList]

Generated Variables
|----|------|
|Name|Label |
|----|------|
|YEAR|<none>|
|----|------|

Processing Statistics
|-------------|-|
|Variables In |9|
|Variables Out|2|
|-------------|-|


.  /**/ LIST /*-*/.
List
|-----------------------------|---------------------------|
|Output Created               |02-APR-2008 16:13:33       |
|-----------------------------|---------------------------|
[YearList]
  id YEAR
[List of cases suppressed]
Number of cases read:  28    Number of cases listed:  28


MATCH FILES
    /FILE=YearList
    /FILE=TestInpt
    /BY ID YEAR.
LIST.

List
|-----------------------------|---------------------------|
|Output Created               |02-APR-2008 16:13:34       |
|-----------------------------|---------------------------|
id YEAR var1 var2

  1 1999    .    .
  1 2000    2    3
  1 2001    5    2
  1 2002    8    2
  1 2003    .    .
  1 2004    .    .
  1 2005    .    .
  2 1999    .    .
  2 2000    2    5
  2 2001    .    .
  2 2002    3    3
  2 2003    8    5
  2 2004    .    .
  2 2005    .    .
  3 1999    .    .
  3 2000    .    .
  3 2001    .    .
  3 2002    .    .
  3 2003    .    .
  3 2004    9    6
  3 2005    .    .
  4 1999    .    .
  4 2000    .    .
  4 2001    .    .
  4 2002    .    .
  4 2003    .    .
  4 2004    .    .
  4 2005    0    2

Number of cases read:  28    Number of cases listed:  28
=================================
APPENDIX: Test data, and all code
(Test data from original posting,
and very nicely done)
=================================
DATA LIST LIST /id year var1 var2.
BEGIN DATA
1 2000 2 3
1 2001 5 2
1 2002 8 2
2 2000 2 5
2 2002 3 3
2 2003 8 5
3 2004 9 6
4 2005 0 2
END DATA.
FORMATS id         (F2)
         year       (F4)
         var1  var2 (F3).
DATASET NAME      TestInpt  WINDOW=FRONT.
LIST.

DATASET DECLARE   YearList.
AGGREGATE OUTFILE=YearList
   /BREAK=ID
   /NRECS 'No. of records for ID, but don''t really care' = NU.
DATASET ACTIVATE  YearList WINDOW=FRONT.

.  /**/ LIST /*-*/.

NUMERIC    YEAR1999 TO YEAR2005 (F4).
DO REPEAT  VARIABLE = YEAR1999 TO YEAR2005
           /VALUE    =     1999 TO     2005.
.  COMPUTE VARIABLE = VALUE.
END REPEAT.

.  /**/ LIST /*-*/.

VARSTOCASES
   /MAKE YEAR FROM YEAR1999 YEAR2000 YEAR2001 YEAR2002 YEAR2003
                   YEAR2004 YEAR2005
  /KEEP =  id
  /NULL = KEEP.

.  /**/ LIST /*-*/.

MATCH FILES
    /FILE=YearList
    /FILE=TestInpt
    /BY ID YEAR.
LIST.

=====================
To manage your subscription to SPSSX-L, send a message to
[hidden email] (not to SPSSX-L), with no body text except the
command. To leave the list, send the command
SIGNOFF SPSSX-L
For a list of commands to manage subscriptions, send the command
INFO REFCARD
Richard Ristow
Reply | Threaded
Open this post in threaded view
♦
♦
|

Re: Adding missing cases to univariate data file

Richard Ristow
In reply to this post by Richard Ristow
At 03:37 PM 4/3/2008, Gross, Paul Jacob wrote (to the list, but I
only received it off-list):

>Richard & SPSS Folks,
>
>Thanks. This worked like a charm. I also found some code that I
>adapted (see below) to do the trick.

Yes: LOOP/XSAVE logic. Another excellent approach.

>For the same data set I would like to insert the missing cases from
>the first year of data [found for the ID] forward to 2005.

Here it is, using a variation on the VARSTOCASES logic; I've noted
the changes, in comments.  "Three things you should be wary of: ... A
code that runs first time", but this one nearly did. Two errors (now
removed), that amounted to typos:


|-----------------------------|---------------------------|
|Output Created               |04-APR-2008 12:17:39       |
|-----------------------------|---------------------------|
[TestInpt]

id year var1 var2

  1 2000    2    3
  1 2001    5    2
  1 2002    8    2
  2 2000    2    5
  2 2002    3    3
  2 2003    8    5
  3 2004    9    6
  4 2005    0    2

Number of cases read:  8    Number of cases listed:  8


DATASET DECLARE   YearList.
*  Change the AGGREGATE: add 'FrstYear'.
AGGREGATE OUTFILE=YearList
   /BREAK=ID
   /NRECS    'No. of records for ID, but don''t really care' = NU
   /FrstYear 'Earliest year found for the ID' = MIN(year).


DATASET ACTIVATE  YearList WINDOW=FRONT.
.  /**/ LIST /*-*/.

List
|-----------------------------|---------------------------|
|Output Created               |04-APR-2008 12:17:40       |
|-----------------------------|---------------------------|
[YearList]

id   NRECS FrstYear

  1       3   2000
  2       3   2000
  3       1   2004
  4       1   2005

Number of cases read:  4    Number of cases listed:  4


NUMERIC    YEAR1999 TO YEAR2005 (F4).
DO REPEAT  VARIABLE = YEAR1999 TO YEAR2005
           /VALUE    =     1999 TO     2005.
*--Statement was:       ....
*  COMPUTE VARIABLE = VALUE.
*--Change to:           ....
.  IF      VALUE GE FrstYear
            VARIABLE = VALUE.
END REPEAT.

.  /**/ LIST /*-*/.

List
|-----------------------------|---------------------------|
|Output Created               |04-APR-2008 12:17:40       |
|-----------------------------|---------------------------|
[YearList]

            Frst
id   NRECS Year YEAR1999 YEAR2000 YEAR2001 YEAR2002 YEAR2003 YEAR2004 YEAR2005

  1       3 2000      .     2000     2001     2002     2003     2004     2005
  2       3 2000      .     2000     2001     2002     2003     2004     2005
  3       1 2004      .        .        .        .        .     2004     2005
  4       1 2005      .        .        .        .        .        .     2005

Number of cases read:  4    Number of cases listed:  4


VARSTOCASES
   /MAKE YEAR FROM YEAR1999 YEAR2000 YEAR2001 YEAR2002 YEAR2003
                   YEAR2004 YEAR2005
  /KEEP =  id
  /NULL =  DROP /* instead of "NULL=KEEP" */.


Variables to Cases
|-----------------------------|---------------------------|
|Output Created               |04-APR-2008 12:17:40       |
|-----------------------------|---------------------------|
[YearList]

Generated Variables
|----|------|
|Name|Label |
|----|------|
|YEAR|<none>|
|----|------|

Processing Statistics
|-------------|--|
|Variables In |10|
|Variables Out|2 |
|-------------|--|


.  /**/ LIST /*-*/.

List
|-----------------------------|---------------------------|
|Output Created               |04-APR-2008 12:17:40       |
|-----------------------------|---------------------------|
[YearList]

id YEAR
[List of records suppressed]
Number of cases read:  15    Number of cases listed:  15


MATCH FILES
    /FILE=YearList
    /FILE=TestInpt
    /BY ID YEAR.
LIST.

List
|-----------------------------|---------------------------|
|Output Created               |04-APR-2008 12:17:41       |
|-----------------------------|---------------------------|
id YEAR var1 var2

  1 2000    2    3
  1 2001    5    2
  1 2002    8    2
  1 2003    .    .
  1 2004    .    .
  1 2005    .    .
  2 2000    2    5
  2 2001    .    .
  2 2002    3    3
  2 2003    8    5
  2 2004    .    .
  2 2005    .    .
  3 2004    9    6
  3 2005    .    .
  4 2005    0    2

Number of cases read:  15    Number of cases listed:  15

============================
APPENDIX: Test data and code
============================
DATA LIST LIST /id year var1 var2.
BEGIN DATA
1 2000 2 3
1 2001 5 2
1 2002 8 2
2 2000 2 5
2 2002 3 3
2 2003 8 5
3 2004 9 6
4 2005 0 2
END DATA.
FORMATS id         (F2)
         year       (F4)
         var1  var2 (F3).
DATASET NAME      TestInpt  WINDOW=FRONT.
LIST.

DATASET DECLARE   YearList.
*  Change the AGGREGATE: add 'FrstYear'.
AGGREGATE OUTFILE=YearList
   /BREAK=ID
   /NRECS    'No. of records for ID, but don''t really care' = NU
   /FrstYear 'Earliest year found for the ID' = MIN(year).


DATASET ACTIVATE  YearList WINDOW=FRONT.

.  /**/ LIST /*-*/.

NUMERIC    YEAR1999 TO YEAR2005 (F4).
DO REPEAT  VARIABLE = YEAR1999 TO YEAR2005
           /VALUE    =     1999 TO     2005.
*--Statement was:       ....
*  COMPUTE VARIABLE = VALUE.
*--Change to:           ....
.  IF      VALUE GE FrstYear
            VARIABLE = VALUE.
END REPEAT.

.  /**/ LIST /*-*/.

VARSTOCASES
   /MAKE YEAR FROM YEAR1999 YEAR2000 YEAR2001 YEAR2002 YEAR2003
                   YEAR2004 YEAR2005
  /KEEP =  id
  /NULL =  DROP /* instead of "NULL=KEEP" */.

.  /**/ LIST /*-*/.

MATCH FILES
    /FILE=YearList
    /FILE=TestInpt
    /BY ID YEAR.
LIST.

=====================
To manage your subscription to SPSSX-L, send a message to
[hidden email] (not to SPSSX-L), with no body text except the
command. To leave the list, send the command
SIGNOFF SPSSX-L
For a list of commands to manage subscriptions, send the command
INFO REFCARD
Jake Gross
Reply | Threaded
Open this post in threaded view
♦
♦
|

Re: Adding missing cases to univariate data file

Jake Gross
Richard,

Thanks once again. This worked quite well, even on a relatively large data
set (about 80,000 cases). You have helped keep my analysis moving along!
Thanks so much.

By the way, my apologies to you and everyone for posting the same question
twice. When I realized it hadn't gone through the first time, I reposted.

Peace,
Jake

-----Original Message-----
From: Richard Ristow [mailto:[hidden email]]
Sent: Friday, April 04, 2008 12:25 PM
To: Gross, Paul Jacob; [hidden email]
Subject: RE: Adding missing cases to univariate data file

At 03:37 PM 4/3/2008, Gross, Paul Jacob wrote (to the list, but I
only received it off-list):

>Richard & SPSS Folks,
>
>Thanks. This worked like a charm. I also found some code that I
>adapted (see below) to do the trick.

Yes: LOOP/XSAVE logic. Another excellent approach.

>For the same data set I would like to insert the missing cases from
>the first year of data [found for the ID] forward to 2005.

Here it is, using a variation on the VARSTOCASES logic; I've noted
the changes, in comments.  "Three things you should be wary of: ... A
code that runs first time", but this one nearly did. Two errors (now
removed), that amounted to typos:


|-----------------------------|---------------------------|
|Output Created               |04-APR-2008 12:17:39       |
|-----------------------------|---------------------------|
[TestInpt]

id year var1 var2

  1 2000    2    3
  1 2001    5    2
  1 2002    8    2
  2 2000    2    5
  2 2002    3    3
  2 2003    8    5
  3 2004    9    6
  4 2005    0    2

Number of cases read:  8    Number of cases listed:  8


DATASET DECLARE   YearList.
*  Change the AGGREGATE: add 'FrstYear'.
AGGREGATE OUTFILE=YearList
   /BREAK=ID
   /NRECS    'No. of records for ID, but don''t really care' = NU
   /FrstYear 'Earliest year found for the ID' = MIN(year).


DATASET ACTIVATE  YearList WINDOW=FRONT.
.  /**/ LIST /*-*/.

List
|-----------------------------|---------------------------|
|Output Created               |04-APR-2008 12:17:40       |
|-----------------------------|---------------------------|
[YearList]

id   NRECS FrstYear

  1       3   2000
  2       3   2000
  3       1   2004
  4       1   2005

Number of cases read:  4    Number of cases listed:  4


NUMERIC    YEAR1999 TO YEAR2005 (F4).
DO REPEAT  VARIABLE = YEAR1999 TO YEAR2005
           /VALUE    =     1999 TO     2005.
*--Statement was:       ....
*  COMPUTE VARIABLE = VALUE.
*--Change to:           ....
.  IF      VALUE GE FrstYear
            VARIABLE = VALUE.
END REPEAT.

.  /**/ LIST /*-*/.

List
|-----------------------------|---------------------------|
|Output Created               |04-APR-2008 12:17:40       |
|-----------------------------|---------------------------|
[YearList]

            Frst
id   NRECS Year YEAR1999 YEAR2000 YEAR2001 YEAR2002 YEAR2003 YEAR2004
YEAR2005

  1       3 2000      .     2000     2001     2002     2003     2004
2005
  2       3 2000      .     2000     2001     2002     2003     2004
2005
  3       1 2004      .        .        .        .        .     2004
2005
  4       1 2005      .        .        .        .        .        .
2005

Number of cases read:  4    Number of cases listed:  4


VARSTOCASES
   /MAKE YEAR FROM YEAR1999 YEAR2000 YEAR2001 YEAR2002 YEAR2003
                   YEAR2004 YEAR2005
  /KEEP =  id
  /NULL =  DROP /* instead of "NULL=KEEP" */.


Variables to Cases
|-----------------------------|---------------------------|
|Output Created               |04-APR-2008 12:17:40       |
|-----------------------------|---------------------------|
[YearList]

Generated Variables
|----|------|
|Name|Label |
|----|------|
|YEAR|<none>|
|----|------|

Processing Statistics
|-------------|--|
|Variables In |10|
|Variables Out|2 |
|-------------|--|


.  /**/ LIST /*-*/.

List
|-----------------------------|---------------------------|
|Output Created               |04-APR-2008 12:17:40       |
|-----------------------------|---------------------------|
[YearList]

id YEAR
[List of records suppressed]
Number of cases read:  15    Number of cases listed:  15


MATCH FILES
    /FILE=YearList
    /FILE=TestInpt
    /BY ID YEAR.
LIST.

List
|-----------------------------|---------------------------|
|Output Created               |04-APR-2008 12:17:41       |
|-----------------------------|---------------------------|
id YEAR var1 var2

  1 2000    2    3
  1 2001    5    2
  1 2002    8    2
  1 2003    .    .
  1 2004    .    .
  1 2005    .    .
  2 2000    2    5
  2 2001    .    .
  2 2002    3    3
  2 2003    8    5
  2 2004    .    .
  2 2005    .    .
  3 2004    9    6
  3 2005    .    .
  4 2005    0    2

Number of cases read:  15    Number of cases listed:  15

============================
APPENDIX: Test data and code
============================
DATA LIST LIST /id year var1 var2.
BEGIN DATA
1 2000 2 3
1 2001 5 2
1 2002 8 2
2 2000 2 5
2 2002 3 3
2 2003 8 5
3 2004 9 6
4 2005 0 2
END DATA.
FORMATS id         (F2)
         year       (F4)
         var1  var2 (F3).
DATASET NAME      TestInpt  WINDOW=FRONT.
LIST.

DATASET DECLARE   YearList.
*  Change the AGGREGATE: add 'FrstYear'.
AGGREGATE OUTFILE=YearList
   /BREAK=ID
   /NRECS    'No. of records for ID, but don''t really care' = NU
   /FrstYear 'Earliest year found for the ID' = MIN(year).


DATASET ACTIVATE  YearList WINDOW=FRONT.

.  /**/ LIST /*-*/.

NUMERIC    YEAR1999 TO YEAR2005 (F4).
DO REPEAT  VARIABLE = YEAR1999 TO YEAR2005
           /VALUE    =     1999 TO     2005.
*--Statement was:       ....
*  COMPUTE VARIABLE = VALUE.
*--Change to:           ....
.  IF      VALUE GE FrstYear
            VARIABLE = VALUE.
END REPEAT.

.  /**/ LIST /*-*/.

VARSTOCASES
   /MAKE YEAR FROM YEAR1999 YEAR2000 YEAR2001 YEAR2002 YEAR2003
                   YEAR2004 YEAR2005
  /KEEP =  id
  /NULL =  DROP /* instead of "NULL=KEEP" */.

.  /**/ LIST /*-*/.

MATCH FILES
    /FILE=YearList
    /FILE=TestInpt
    /BY ID YEAR.
LIST.

=====================
To manage your subscription to SPSSX-L, send a message to
[hidden email] (not to SPSSX-L), with no body text except the
command. To leave the list, send the command
SIGNOFF SPSSX-L
For a list of commands to manage subscriptions, send the command
INFO REFCARD
Free forum by Nabble Edit this page