SPSSX Discussion

Ranking within cases individually

Classic

List

Threaded

15 messages Options

Erin Miles

Ranking within cases individually

I have a data set that gives me plant percent cover data for 2100 plots
(rows) and there are about 250 potential species that can be present in each
plot (rows). I need to sort/rank each row so that the percent covers go
from largest to smallest. For this purpose I do not need to retain the
species name, I just need to rank the percent covers from greatest to least
for EACH plot in order to create Species Rank-Abundance graphs. Can anyone
help me with this task? Much Thanks! Erin

_________________________________________________________________
Get free, personalized commercial-free online radio with MSN Radio powered
by Pandora http://radio.msn.com/?icid=T002MSN03A07001

Maguin, Eugene

Re: Ranking within cases individually

Erin,

I'm probably not understanding what you need to do but why doesn't the Sort
cases command do what you need? Is there something about the structure of
your dataset that prevents you from using this command to get what you need?

Gene Maguin

Richard Ristow

Re: Ranking within cases individually

In reply to this post by Erin Miles

(Marta and Jan, I'm copying to you because you're probably the best
MATRIX whizzes among us.)

At 03:00 PM 12/8/2006, Erin Miles wrote:

>I have a data set that gives me plant percent cover data for 2100
>plots (rows) and there are about 250 potential species that can be
>present in each plot (rows). I need to sort/rank each row so that the
>percent covers go from largest to smallest.

For sorting variables within cases, the best technique that seems to
exist is to use the GRADE function in the MATRIX language, which
apparently calls an internal sort routine; this has been noted as the
Ristow-Peck algorithm. Below is an example, SPSS draft output. Only
just out of testing. There doesn't seem to be a way to do it without a
LOOP in MATRIX.

There's a lot of matrix tracing that could be omitted in production.

The MATRIX SAVE wipes the working file. To join to the pre-existing
file, save the latter and re-attach using MATCH FILES. And, here goes:

LIST.

Notes
|-----------------------------|---------------------------|
|Output Created |11-DEC-2006 13:03:27 |
|-----------------------------|---------------------------|
PLOT_ID COVR01 COVR02 COVR03 COVR04 COVR05

PLOT001 7% 26% 44% 16% 8%
PLOT002 51% 18% 25% 2% 5%
PLOT003 6% 2% 38% 6% 48%
PLOT004 51% 17% 6% 25% 1%
PLOT005 4% 46% 9% 9% 33%
PLOT006 60% 5% 8% 1% 26%
PLOT007 8% 4% 13% 6% 69%
PLOT008 21% 36% 8% 12% 22%
PLOT009 6% 41% 2% 21% 30%
PLOT010 50% 7% 14% 27% 2%

Number of cases read: 10 Number of cases listed: 10

MATRIX.
GET PLOT_ID /FILE=*/VARIABLES=PLOT_ID.
GET COVER /FILE=*/VARIABLES=COVR01 TO COVR05.
PRINT PLOT_ID /FORMAT='A8'.
PRINT COVER/FORMAT='F5.1'/RNAMES=PLOT_ID.

COMPUTE CVR_SORT = 0*COVER.
COMPUTE CVR_ORDR = 0*COVER.

LOOP ROW = 1 TO NROW(COVER).
PRINT ROW.
COMPUTE ROW_ORDR = GRADE(-1*COVER(ROW,:)).
COMPUTE CVR_ORDR(ROW,:) = Row_ORDR.
COMPUTE CVR_SORT(ROW,Row_Ordr) = COVER(Row,:).

COMPUTE ROW_RSLT ={COVER(ROW,:) ;
CVR_ORDR(ROW,:);
CVR_SORT(ROW,:)}.
PRINT ROW_RSLT /FORMAT=F5.1
/RLABELS='Input','Order','Sorted'.

END LOOP.

PRINT CVR_ORDR/RNAMES=PLOT_ID.
PRINT CVR_SORT/FORMAT='F5.1'/RNAMES=PLOT_ID.

SAVE {PLOT_ID,CVR_SORT,CVR_ORDR,COVER}
/OUTFILE=*
/VARIABLES=PLOT_ID,SORTED01 TO SORTED05,
ORDER01 TO ORDER05,
COVER01 To COVER05
/STRINGS=PLOT_ID.
END MATRIX.

Notes
|--------------------------|---------------------------|
|Output Created |11-DEC-2006 13:03:27 |
|--------------------------|---------------------------|
Run MATRIX procedure:

PLOT_ID
PLOT001
PLOT002
PLOT003
PLOT004
PLOT005
PLOT006
PLOT007
PLOT008
PLOT009
PLOT010

COVER
PLOT001 6.9 26.0 43.6 15.8 7.8
PLOT002 50.5 17.9 24.8 2.3 4.6
PLOT003 5.9 1.7 37.5 6.4 48.5
PLOT004 51.4 16.8 5.9 25.2 .7
PLOT005 3.7 45.9 8.9 8.7 32.8
PLOT006 59.7 5.2 8.0 1.1 26.0
PLOT007 7.8 3.7 12.8 6.4 69.2
PLOT008 21.4 36.5 7.8 12.4 21.9
PLOT009 6.4 41.0 1.9 20.7 30.0
PLOT010 50.4 7.0 13.6 26.6 2.5

ROW
1

ROW_RSLT
Input 6.9 26.0 43.6 15.8 7.8
Order 5.0 2.0 1.0 3.0 4.0
Sorted 43.6 26.0 15.8 7.8 6.9

ROW
2

ROW_RSLT
Input 50.5 17.9 24.8 2.3 4.6
Order 1.0 3.0 2.0 5.0 4.0
Sorted 50.5 24.8 17.9 4.6 2.3

ROW
3

ROW_RSLT
Input 5.9 1.7 37.5 6.4 48.5
Order 4.0 5.0 2.0 3.0 1.0
Sorted 48.5 37.5 6.4 5.9 1.7

ROW
4

ROW_RSLT
Input 51.4 16.8 5.9 25.2 .7
Order 1.0 3.0 4.0 2.0 5.0
Sorted 51.4 25.2 16.8 5.9 .7

ROW
5

ROW_RSLT
Input 3.7 45.9 8.9 8.7 32.8
Order 5.0 1.0 3.0 4.0 2.0
Sorted 45.9 32.8 8.9 8.7 3.7

ROW
6

ROW_RSLT
Input 59.7 5.2 8.0 1.1 26.0
Order 1.0 4.0 3.0 5.0 2.0
Sorted 59.7 26.0 8.0 5.2 1.1

ROW
7

ROW_RSLT
Input 7.8 3.7 12.8 6.4 69.2
Order 3.0 5.0 2.0 4.0 1.0
Sorted 69.2 12.8 7.8 6.4 3.7

ROW
8

ROW_RSLT
Input 21.4 36.5 7.8 12.4 21.9
Order 3.0 1.0 5.0 4.0 2.0
Sorted 36.5 21.9 21.4 12.4 7.8

ROW
9

ROW_RSLT
Input 6.4 41.0 1.9 20.7 30.0
Order 4.0 1.0 5.0 3.0 2.0
Sorted 41.0 30.0 20.7 6.4 1.9

ROW
10

ROW_RSLT
Input 50.4 7.0 13.6 26.6 2.5
Order 1.0 4.0 3.0 2.0 5.0
Sorted 50.4 26.6 13.6 7.0 2.5

CVR_ORDR
PLOT001 5 2 1 3 4
PLOT002 1 3 2 5 4
PLOT003 4 5 2 3 1
PLOT004 1 3 4 2 5
PLOT005 5 1 3 4 2
PLOT006 1 4 3 5 2
PLOT007 3 5 2 4 1
PLOT008 3 1 5 4 2
PLOT009 4 1 5 3 2
PLOT010 1 4 3 2 5

CVR_SORT
PLOT001 43.6 26.0 15.8 7.8 6.9
PLOT002 50.5 24.8 17.9 4.6 2.3
PLOT003 48.5 37.5 6.4 5.9 1.7
PLOT004 51.4 25.2 16.8 5.9 .7
PLOT005 45.9 32.8 8.9 8.7 3.7
PLOT006 59.7 26.0 8.0 5.2 1.1
PLOT007 69.2 12.8 7.8 6.4 3.7
PLOT008 36.5 21.9 21.4 12.4 7.8
PLOT009 41.0 30.0 20.7 6.4 1.9
PLOT010 50.4 26.6 13.6 7.0 2.5

------ END MATRIX -----

FORMATS ORDER01 TO ORDER05 (F3)
/SORTED01 TO SORTED05 (PCT3),
/COVER01 TO COVER05 (PCT3).
TEMPORARY.
STRING SPACE(A8).
LIST VARIABLES=PLOT_ID
SORTED01 TO SORTED05
ORDER01 TO ORDER05
COVER01 TO COVER05
SPACE.

Notes
|-----------------------------|---------------------------|
|Output Created |11-DEC-2006 13:03:30 |
|-----------------------------|---------------------------|
SOR SOR SOR SOR SOR ORD ORD ORD ORD ORD COV COV
TED TED TED TED TED ER0 ER0 ER0 ER0 ER0 ER0 ER0 COVE COVE
COVE
PLOT_ID 01 02 03 04 05 1 2 3 4 5 1 2 R03 R04 R05
SPACE

PLOT001 44% 26% 16% 8% 7% 5 2 1 3 4 7% 26% 44% 16% 8%
PLOT002 51% 25% 18% 5% 2% 1 3 2 5 4 51% 18% 25% 2% 5%
PLOT003 48% 38% 6% 6% 2% 4 5 2 3 1 6% 2% 38% 6% 48%
PLOT004 51% 25% 17% 6% 1% 1 3 4 2 5 51% 17% 6% 25% 1%
PLOT005 46% 33% 9% 9% 4% 5 1 3 4 2 4% 46% 9% 9% 33%
PLOT006 60% 26% 8% 5% 1% 1 4 3 5 2 60% 5% 8% 1% 26%
PLOT007 69% 13% 8% 6% 4% 3 5 2 4 1 8% 4% 13% 6% 69%
PLOT008 36% 22% 21% 12% 8% 3 1 5 4 2 21% 36% 8% 12% 22%
PLOT009 41% 30% 21% 6% 2% 4 1 5 3 2 6% 41% 2% 21% 30%
PLOT010 50% 27% 14% 7% 2% 1 4 3 2 5 50% 7% 14% 27% 2%

Number of cases read: 10 Number of cases listed: 10

===================
APPENDIX: TEST DATA
===================
* ............ Test data ............ .
* 05 species, 10 cases (plots) .
* Abundances exponentially distributed with mean 100, .
* normalized to total abundance 1 and converted to (UGH!) .
* SPSS 'percents' for display. .

NEW FILE.
INPUT PROGRAM.
. STRING PLOT_ID (A8).
. NUMERIC COVR01 TO COVR05 (F4).
. VECTOR COVER =
COVR01 TO COVR05.
. LOOP #PLOT = 1 TO 10.
. COMPUTE PLOT_ID = CONCAT('PLOT',STRING(#PLOT,N3)).
. LOOP #PLANT = 1 TO 5.
. COMPUTE COVER(#PLANT)=RV.EXP(.01).
. END LOOP.
* Normalize, to total cover = 1 . .
. COMPUTE #TOTAL = SUM(COVR01 TO COVR05).
. LOOP #PLANT = 1 TO 05.
* The multiplication by 100, below, is because of .
* SPSS's very unfortunate implementation of PCT .
* formats. Sigh. .
. COMPUTE COVER(#PLANT)=100*COVER(#PLANT)/#TOTAL.
. END LOOP.
. END CASE.
. END LOOP.
. FORMATS COVR01 TO COVR05 (PCT3).
END FILE.
END INPUT PROGRAM.

Peck, Jon

Re: Ranking within cases individually

Here is a Python solution assuming SPSS 15 and programmability are installed.
First it defines a general function for this task and then it calls it with two variables as an example.
The part in triple quotes is the documentation for the function. sysmis values sort low.
It creates new variables with the sorted values.

The "for v in varlist" loop creates new variables; the "for case in curs" loops over the cases

begin program.
import spss, spssdata

def vsort(varlist, direction='a', suffix='_sorted'):
"""sort variables, assumed to be numeric, within cases creating new variables

varlist is the list of variables to sort.
direction can be 'a' or 'd' for ascending or descending.
suffix is the string to append to the original name for the
corresponding new variable."""

rev = direction =='d'
curs = spssdata.Spssdata(indexes=varlist, names=False, accessType='w')
for v in varlist:
curs.append(v + suffix)
curs.commitdict()
for case in curs:
curs.casevalues(sorted(case, reverse = rev))
curs.Cclose()

vsort(['salbegin','salary'])
end program.

-Jon Peck

-----Original Message-----
From: SPSSX(r) Discussion [mailto:[hidden email]] On Behalf Of Richard Ristow
Sent: Monday, December 11, 2006 12:14 PM
To: [hidden email]
Subject: Re: [SPSSX-L] Ranking within cases individually

(Marta and Jan, I'm copying to you because you're probably the best
MATRIX whizzes among us.)

At 03:00 PM 12/8/2006, Erin Miles wrote:

>I have a data set that gives me plant percent cover data for 2100
>plots (rows) and there are about 250 potential species that can be
>present in each plot (rows). I need to sort/rank each row so that the
>percent covers go from largest to smallest.

For sorting variables within cases, the best technique that seems to
exist is to use the GRADE function in the MATRIX language, which
apparently calls an internal sort routine; this has been noted as the
Ristow-Peck algorithm. Below is an example, SPSS draft output. Only
just out of testing. There doesn't seem to be a way to do it without a
LOOP in MATRIX.

There's a lot of matrix tracing that could be omitted in production.

The MATRIX SAVE wipes the working file. To join to the pre-existing
file, save the latter and re-attach using MATCH FILES. And, here goes:

LIST.

Notes
|-----------------------------|---------------------------|
|Output Created |11-DEC-2006 13:03:27 |
|-----------------------------|---------------------------|
PLOT_ID COVR01 COVR02 COVR03 COVR04 COVR05

PLOT001 7% 26% 44% 16% 8%
PLOT002 51% 18% 25% 2% 5%
PLOT003 6% 2% 38% 6% 48%
PLOT004 51% 17% 6% 25% 1%
PLOT005 4% 46% 9% 9% 33%
PLOT006 60% 5% 8% 1% 26%
PLOT007 8% 4% 13% 6% 69%
PLOT008 21% 36% 8% 12% 22%
PLOT009 6% 41% 2% 21% 30%
PLOT010 50% 7% 14% 27% 2%

Number of cases read: 10 Number of cases listed: 10

MATRIX.
GET PLOT_ID /FILE=*/VARIABLES=PLOT_ID.
GET COVER /FILE=*/VARIABLES=COVR01 TO COVR05.
PRINT PLOT_ID /FORMAT='A8'.
PRINT COVER/FORMAT='F5.1'/RNAMES=PLOT_ID.

COMPUTE CVR_SORT = 0*COVER.
COMPUTE CVR_ORDR = 0*COVER.

LOOP ROW = 1 TO NROW(COVER).
PRINT ROW.
COMPUTE ROW_ORDR = GRADE(-1*COVER(ROW,:)).
COMPUTE CVR_ORDR(ROW,:) = Row_ORDR.
COMPUTE CVR_SORT(ROW,Row_Ordr) = COVER(Row,:).

COMPUTE ROW_RSLT ={COVER(ROW,:) ;
CVR_ORDR(ROW,:);
CVR_SORT(ROW,:)}.
PRINT ROW_RSLT /FORMAT=F5.1
/RLABELS='Input','Order','Sorted'.

END LOOP.

PRINT CVR_ORDR/RNAMES=PLOT_ID.
PRINT CVR_SORT/FORMAT='F5.1'/RNAMES=PLOT_ID.

SAVE {PLOT_ID,CVR_SORT,CVR_ORDR,COVER}
/OUTFILE=*
/VARIABLES=PLOT_ID,SORTED01 TO SORTED05,
ORDER01 TO ORDER05,
COVER01 To COVER05
/STRINGS=PLOT_ID.
END MATRIX.

Notes
|--------------------------|---------------------------|
|Output Created |11-DEC-2006 13:03:27 |
|--------------------------|---------------------------|
Run MATRIX procedure:

PLOT_ID
PLOT001
PLOT002
PLOT003
PLOT004
PLOT005
PLOT006
PLOT007
PLOT008
PLOT009
PLOT010

COVER
PLOT001 6.9 26.0 43.6 15.8 7.8
PLOT002 50.5 17.9 24.8 2.3 4.6
PLOT003 5.9 1.7 37.5 6.4 48.5
PLOT004 51.4 16.8 5.9 25.2 .7
PLOT005 3.7 45.9 8.9 8.7 32.8
PLOT006 59.7 5.2 8.0 1.1 26.0
PLOT007 7.8 3.7 12.8 6.4 69.2
PLOT008 21.4 36.5 7.8 12.4 21.9
PLOT009 6.4 41.0 1.9 20.7 30.0
PLOT010 50.4 7.0 13.6 26.6 2.5

ROW
1

ROW_RSLT
Input 6.9 26.0 43.6 15.8 7.8
Order 5.0 2.0 1.0 3.0 4.0
Sorted 43.6 26.0 15.8 7.8 6.9

ROW
2

ROW_RSLT
Input 50.5 17.9 24.8 2.3 4.6
Order 1.0 3.0 2.0 5.0 4.0
Sorted 50.5 24.8 17.9 4.6 2.3

ROW
3

ROW_RSLT
Input 5.9 1.7 37.5 6.4 48.5
Order 4.0 5.0 2.0 3.0 1.0
Sorted 48.5 37.5 6.4 5.9 1.7

ROW
4

ROW_RSLT
Input 51.4 16.8 5.9 25.2 .7
Order 1.0 3.0 4.0 2.0 5.0
Sorted 51.4 25.2 16.8 5.9 .7

ROW
5

ROW_RSLT
Input 3.7 45.9 8.9 8.7 32.8
Order 5.0 1.0 3.0 4.0 2.0
Sorted 45.9 32.8 8.9 8.7 3.7

ROW
6

ROW_RSLT
Input 59.7 5.2 8.0 1.1 26.0
Order 1.0 4.0 3.0 5.0 2.0
Sorted 59.7 26.0 8.0 5.2 1.1

ROW
7

ROW_RSLT
Input 7.8 3.7 12.8 6.4 69.2
Order 3.0 5.0 2.0 4.0 1.0
Sorted 69.2 12.8 7.8 6.4 3.7

ROW
8

ROW_RSLT
Input 21.4 36.5 7.8 12.4 21.9
Order 3.0 1.0 5.0 4.0 2.0
Sorted 36.5 21.9 21.4 12.4 7.8

ROW
9

ROW_RSLT
Input 6.4 41.0 1.9 20.7 30.0
Order 4.0 1.0 5.0 3.0 2.0
Sorted 41.0 30.0 20.7 6.4 1.9

ROW
10

ROW_RSLT
Input 50.4 7.0 13.6 26.6 2.5
Order 1.0 4.0 3.0 2.0 5.0
Sorted 50.4 26.6 13.6 7.0 2.5

CVR_ORDR
PLOT001 5 2 1 3 4
PLOT002 1 3 2 5 4
PLOT003 4 5 2 3 1
PLOT004 1 3 4 2 5
PLOT005 5 1 3 4 2
PLOT006 1 4 3 5 2
PLOT007 3 5 2 4 1
PLOT008 3 1 5 4 2
PLOT009 4 1 5 3 2
PLOT010 1 4 3 2 5

CVR_SORT
PLOT001 43.6 26.0 15.8 7.8 6.9
PLOT002 50.5 24.8 17.9 4.6 2.3
PLOT003 48.5 37.5 6.4 5.9 1.7
PLOT004 51.4 25.2 16.8 5.9 .7
PLOT005 45.9 32.8 8.9 8.7 3.7
PLOT006 59.7 26.0 8.0 5.2 1.1
PLOT007 69.2 12.8 7.8 6.4 3.7
PLOT008 36.5 21.9 21.4 12.4 7.8
PLOT009 41.0 30.0 20.7 6.4 1.9
PLOT010 50.4 26.6 13.6 7.0 2.5

------ END MATRIX -----

FORMATS ORDER01 TO ORDER05 (F3)
/SORTED01 TO SORTED05 (PCT3),
/COVER01 TO COVER05 (PCT3).
TEMPORARY.
STRING SPACE(A8).
LIST VARIABLES=PLOT_ID
SORTED01 TO SORTED05
ORDER01 TO ORDER05
COVER01 TO COVER05
SPACE.

Notes
|-----------------------------|---------------------------|
|Output Created |11-DEC-2006 13:03:30 |
|-----------------------------|---------------------------|
SOR SOR SOR SOR SOR ORD ORD ORD ORD ORD COV COV
TED TED TED TED TED ER0 ER0 ER0 ER0 ER0 ER0 ER0 COVE COVE
COVE
PLOT_ID 01 02 03 04 05 1 2 3 4 5 1 2 R03 R04 R05
SPACE

PLOT001 44% 26% 16% 8% 7% 5 2 1 3 4 7% 26% 44% 16% 8%
PLOT002 51% 25% 18% 5% 2% 1 3 2 5 4 51% 18% 25% 2% 5%
PLOT003 48% 38% 6% 6% 2% 4 5 2 3 1 6% 2% 38% 6% 48%
PLOT004 51% 25% 17% 6% 1% 1 3 4 2 5 51% 17% 6% 25% 1%
PLOT005 46% 33% 9% 9% 4% 5 1 3 4 2 4% 46% 9% 9% 33%
PLOT006 60% 26% 8% 5% 1% 1 4 3 5 2 60% 5% 8% 1% 26%
PLOT007 69% 13% 8% 6% 4% 3 5 2 4 1 8% 4% 13% 6% 69%
PLOT008 36% 22% 21% 12% 8% 3 1 5 4 2 21% 36% 8% 12% 22%
PLOT009 41% 30% 21% 6% 2% 4 1 5 3 2 6% 41% 2% 21% 30%
PLOT010 50% 27% 14% 7% 2% 1 4 3 2 5 50% 7% 14% 27% 2%

Number of cases read: 10 Number of cases listed: 10

===================
APPENDIX: TEST DATA
===================
* ............ Test data ............ .
* 05 species, 10 cases (plots) .
* Abundances exponentially distributed with mean 100, .
* normalized to total abundance 1 and converted to (UGH!) .
* SPSS 'percents' for display. .

NEW FILE.
INPUT PROGRAM.
. STRING PLOT_ID (A8).
. NUMERIC COVR01 TO COVR05 (F4).
. VECTOR COVER =
COVR01 TO COVR05.
. LOOP #PLOT = 1 TO 10.
. COMPUTE PLOT_ID = CONCAT('PLOT',STRING(#PLOT,N3)).
. LOOP #PLANT = 1 TO 5.
. COMPUTE COVER(#PLANT)=RV.EXP(.01).
. END LOOP.
* Normalize, to total cover = 1 . .
. COMPUTE #TOTAL = SUM(COVR01 TO COVR05).
. LOOP #PLANT = 1 TO 05.
* The multiplication by 100, below, is because of .
* SPSS's very unfortunate implementation of PCT .
* formats. Sigh. .
. COMPUTE COVER(#PLANT)=100*COVER(#PLANT)/#TOTAL.
. END LOOP.
. END CASE.
. END LOOP.
. FORMATS COVR01 TO COVR05 (PCT3).
END FILE.
END INPUT PROGRAM.

Dennis Deck

Re: Ranking within cases individually

In reply to this post by Erin Miles

Assuming that each observation represents a separate record something
like:

ObservationID PlotNumber SpeciesCode PctCover

Then:

SORT CASES BY PlotNumber PctCover (D) .

Puts records in descending order of PctCover within Plot

Alternatively you could use split file processing with RANK:

SORT CASES BY Plot
SPLIT FILE BY Plot
RANK Vars= PctCover D .

Check RANK syntax for desired subcommands.

Dennis Deck, PhD
RMC Research Corporation
[hidden email]
-----Original Message-----
From: Erin Miles [mailto:[hidden email]]
Sent: Friday, December 08, 2006 12:01 PM
Subject: Ranking within cases individually

I have a data set that gives me plant percent cover data for 2100 plots
(rows) and there are about 250 potential species that can be present in
each
plot (rows). I need to sort/rank each row so that the percent covers go
from largest to smallest. For this purpose I do not need to retain the
species name, I just need to rank the percent covers from greatest to
least
for EACH plot in order to create Species Rank-Abundance graphs. Can
anyone
help me with this task? Much Thanks! Erin

_________________________________________________________________
Get free, personalized commercial-free online radio with MSN Radio
powered
by Pandora http://radio.msn.com/?icid=T002MSN03A07001

Richard Ristow

Re: Ranking within cases individually

In reply to this post by Erin Miles

A postscript:

First, the other approach I'd consider is to 'unroll' the data with
VARSTOCASES, so each abundance value is in a separate case, and use
SORT CASES.

Second, I've tried out a variation on the code I posted. Here,
"Sourcexx" gives which plant is in the position in the sorted sequence;
formerly, "ORDERxx" gave which position in the sorted sequence the
original plant would have. Here's the output; I won't post the code,
now.

List
|-----------------------------|---------------------------|
|Output Created |11-DEC-2006 14:49:51 |
|-----------------------------|---------------------------|
SOR SOR SOR SOR SOR Sou Sou Sou Sou Sou COV COV COV COV
TED TED TED TED TED rce rce rce rce rce ER0 ER0 ER0 ER0 COVE
PLOT_ID 01 02 03 04 05 01 02 03 04 05 1 2 3 4 R05

PLOT001 57% 18% 11% 7% 7% 4 5 1 3 2 11% 7% 7% 57% 18%
PLOT002 67% 14% 14% 5% 0% 3 1 5 2 4 14% 5% 67% 0% 14%
PLOT003 38% 23% 17% 16% 7% 1 3 5 4 2 38% 7% 23% 16% 17%
PLOT004 50% 24% 11% 9% 6% 5 1 3 2 4 24% 9% 11% 6% 50%
PLOT005 40% 27% 23% 7% 3% 4 5 2 1 3 7% 23% 3% 40% 27%
PLOT006 34% 24% 23% 15% 5% 3 4 2 1 5 15% 23% 34% 24% 5%
PLOT007 49% 20% 18% 13% 0% 4 5 2 3 1 0% 18% 13% 49% 20%
PLOT008 34% 19% 16% 16% 15% 4 5 2 3 1 15% 16% 16% 34% 19%
PLOT009 49% 25% 20% 4% 2% 2 4 1 5 3 20% 49% 2% 25% 4%
PLOT010 62% 14% 13% 10% 1% 5 4 2 3 1 1% 13% 10% 14% 62%

Number of cases read: 10 Number of cases listed: 10

Roberts, Michael

Re: Ranking within cases individually

In reply to this post by Peck, Jon

Jon,

What would I have to do to make this work in version 14.0.1? I tried
running the code but 'accessType' appears to generate a problem!

Thank You

Mike

-----Original Message-----
From: SPSSX(r) Discussion [mailto:[hidden email]] On Behalf Of
Peck, Jon
Sent: Monday, December 11, 2006 1:52 PM
To: [hidden email]
Subject: Re: Ranking within cases individually

Here is a Python solution assuming SPSS 15 and programmability are
installed.
First it defines a general function for this task and then it calls it
with two variables as an example.
The part in triple quotes is the documentation for the function. sysmis
values sort low.
It creates new variables with the sorted values.

The "for v in varlist" loop creates new variables; the "for case in
curs" loops over the cases

begin program.
import spss, spssdata

def vsort(varlist, direction='a', suffix='_sorted'):
"""sort variables, assumed to be numeric, within cases creating
new variables

varlist is the list of variables to sort.
direction can be 'a' or 'd' for ascending or descending.
suffix is the string to append to the original name for the
corresponding new variable."""

rev = direction =='d'
curs = spssdata.Spssdata(indexes=varlist, names=False,
accessType='w')
for v in varlist:
curs.append(v + suffix)
curs.commitdict()
for case in curs:
curs.casevalues(sorted(case, reverse = rev))
curs.Cclose()

vsort(['salbegin','salary'])
end program.

-Jon Peck

-----Original Message-----
From: SPSSX(r) Discussion [mailto:[hidden email]] On Behalf Of
Richard Ristow
Sent: Monday, December 11, 2006 12:14 PM
To: [hidden email]
Subject: Re: [SPSSX-L] Ranking within cases individually

(Marta and Jan, I'm copying to you because you're probably the best
MATRIX whizzes among us.)

At 03:00 PM 12/8/2006, Erin Miles wrote:

>I have a data set that gives me plant percent cover data for 2100 plots

>(rows) and there are about 250 potential species that can be present in

>each plot (rows). I need to sort/rank each row so that the percent
>covers go from largest to smallest.

For sorting variables within cases, the best technique that seems to
exist is to use the GRADE function in the MATRIX language, which
apparently calls an internal sort routine; this has been noted as the
Ristow-Peck algorithm. Below is an example, SPSS draft output. Only just
out of testing. There doesn't seem to be a way to do it without a LOOP
in MATRIX.

There's a lot of matrix tracing that could be omitted in production.

The MATRIX SAVE wipes the working file. To join to the pre-existing
file, save the latter and re-attach using MATCH FILES. And, here goes:

LIST.

Notes
|-----------------------------|---------------------------|
|Output Created |11-DEC-2006 13:03:27 |
|-----------------------------|---------------------------|
PLOT_ID COVR01 COVR02 COVR03 COVR04 COVR05

PLOT001 7% 26% 44% 16% 8%
PLOT002 51% 18% 25% 2% 5%
PLOT003 6% 2% 38% 6% 48%
PLOT004 51% 17% 6% 25% 1%
PLOT005 4% 46% 9% 9% 33%
PLOT006 60% 5% 8% 1% 26%
PLOT007 8% 4% 13% 6% 69%
PLOT008 21% 36% 8% 12% 22%
PLOT009 6% 41% 2% 21% 30%
PLOT010 50% 7% 14% 27% 2%

Number of cases read: 10 Number of cases listed: 10

MATRIX.
GET PLOT_ID /FILE=*/VARIABLES=PLOT_ID.
GET COVER /FILE=*/VARIABLES=COVR01 TO COVR05.
PRINT PLOT_ID /FORMAT='A8'.
PRINT COVER/FORMAT='F5.1'/RNAMES=PLOT_ID.

COMPUTE CVR_SORT = 0*COVER.
COMPUTE CVR_ORDR = 0*COVER.

LOOP ROW = 1 TO NROW(COVER).
PRINT ROW.
COMPUTE ROW_ORDR = GRADE(-1*COVER(ROW,:)).
COMPUTE CVR_ORDR(ROW,:) = Row_ORDR.
COMPUTE CVR_SORT(ROW,Row_Ordr) = COVER(Row,:).

COMPUTE ROW_RSLT ={COVER(ROW,:) ;
CVR_ORDR(ROW,:);
CVR_SORT(ROW,:)}.
PRINT ROW_RSLT /FORMAT=F5.1
/RLABELS='Input','Order','Sorted'.

END LOOP.

PRINT CVR_ORDR/RNAMES=PLOT_ID.
PRINT CVR_SORT/FORMAT='F5.1'/RNAMES=PLOT_ID.

SAVE {PLOT_ID,CVR_SORT,CVR_ORDR,COVER}
/OUTFILE=*
/VARIABLES=PLOT_ID,SORTED01 TO SORTED05,
ORDER01 TO ORDER05,
COVER01 To COVER05
/STRINGS=PLOT_ID.
END MATRIX.

Notes
|--------------------------|---------------------------|
|Output Created |11-DEC-2006 13:03:27 |
|--------------------------|---------------------------|
Run MATRIX procedure:

PLOT_ID
PLOT001
PLOT002
PLOT003
PLOT004
PLOT005
PLOT006
PLOT007
PLOT008
PLOT009
PLOT010

COVER
PLOT001 6.9 26.0 43.6 15.8 7.8
PLOT002 50.5 17.9 24.8 2.3 4.6
PLOT003 5.9 1.7 37.5 6.4 48.5
PLOT004 51.4 16.8 5.9 25.2 .7
PLOT005 3.7 45.9 8.9 8.7 32.8
PLOT006 59.7 5.2 8.0 1.1 26.0
PLOT007 7.8 3.7 12.8 6.4 69.2
PLOT008 21.4 36.5 7.8 12.4 21.9
PLOT009 6.4 41.0 1.9 20.7 30.0
PLOT010 50.4 7.0 13.6 26.6 2.5

ROW
1

ROW_RSLT
Input 6.9 26.0 43.6 15.8 7.8
Order 5.0 2.0 1.0 3.0 4.0
Sorted 43.6 26.0 15.8 7.8 6.9

ROW
2

ROW_RSLT
Input 50.5 17.9 24.8 2.3 4.6
Order 1.0 3.0 2.0 5.0 4.0
Sorted 50.5 24.8 17.9 4.6 2.3

ROW
3

ROW_RSLT
Input 5.9 1.7 37.5 6.4 48.5
Order 4.0 5.0 2.0 3.0 1.0
Sorted 48.5 37.5 6.4 5.9 1.7

ROW
4

ROW_RSLT
Input 51.4 16.8 5.9 25.2 .7
Order 1.0 3.0 4.0 2.0 5.0
Sorted 51.4 25.2 16.8 5.9 .7

ROW
5

ROW_RSLT
Input 3.7 45.9 8.9 8.7 32.8
Order 5.0 1.0 3.0 4.0 2.0
Sorted 45.9 32.8 8.9 8.7 3.7

ROW
6

ROW_RSLT
Input 59.7 5.2 8.0 1.1 26.0
Order 1.0 4.0 3.0 5.0 2.0
Sorted 59.7 26.0 8.0 5.2 1.1

ROW
7

ROW_RSLT
Input 7.8 3.7 12.8 6.4 69.2
Order 3.0 5.0 2.0 4.0 1.0
Sorted 69.2 12.8 7.8 6.4 3.7

ROW
8

ROW_RSLT
Input 21.4 36.5 7.8 12.4 21.9
Order 3.0 1.0 5.0 4.0 2.0
Sorted 36.5 21.9 21.4 12.4 7.8

ROW
9

ROW_RSLT
Input 6.4 41.0 1.9 20.7 30.0
Order 4.0 1.0 5.0 3.0 2.0
Sorted 41.0 30.0 20.7 6.4 1.9

ROW
10

ROW_RSLT
Input 50.4 7.0 13.6 26.6 2.5
Order 1.0 4.0 3.0 2.0 5.0
Sorted 50.4 26.6 13.6 7.0 2.5

CVR_ORDR
PLOT001 5 2 1 3 4
PLOT002 1 3 2 5 4
PLOT003 4 5 2 3 1
PLOT004 1 3 4 2 5
PLOT005 5 1 3 4 2
PLOT006 1 4 3 5 2
PLOT007 3 5 2 4 1
PLOT008 3 1 5 4 2
PLOT009 4 1 5 3 2
PLOT010 1 4 3 2 5

CVR_SORT
PLOT001 43.6 26.0 15.8 7.8 6.9
PLOT002 50.5 24.8 17.9 4.6 2.3
PLOT003 48.5 37.5 6.4 5.9 1.7
PLOT004 51.4 25.2 16.8 5.9 .7
PLOT005 45.9 32.8 8.9 8.7 3.7
PLOT006 59.7 26.0 8.0 5.2 1.1
PLOT007 69.2 12.8 7.8 6.4 3.7
PLOT008 36.5 21.9 21.4 12.4 7.8
PLOT009 41.0 30.0 20.7 6.4 1.9
PLOT010 50.4 26.6 13.6 7.0 2.5

------ END MATRIX -----

FORMATS ORDER01 TO ORDER05 (F3)
/SORTED01 TO SORTED05 (PCT3),
/COVER01 TO COVER05 (PCT3).
TEMPORARY.
STRING SPACE(A8).
LIST VARIABLES=PLOT_ID
SORTED01 TO SORTED05
ORDER01 TO ORDER05
COVER01 TO COVER05
SPACE.

Notes
|-----------------------------|---------------------------|
|Output Created |11-DEC-2006 13:03:30 |
|-----------------------------|---------------------------|
SOR SOR SOR SOR SOR ORD ORD ORD ORD ORD COV COV
TED TED TED TED TED ER0 ER0 ER0 ER0 ER0 ER0 ER0 COVE COVE COVE
PLOT_ID 01 02 03 04 05 1 2 3 4 5 1 2 R03 R04 R05
SPACE

PLOT001 44% 26% 16% 8% 7% 5 2 1 3 4 7% 26% 44% 16% 8%
PLOT002 51% 25% 18% 5% 2% 1 3 2 5 4 51% 18% 25% 2% 5%
PLOT003 48% 38% 6% 6% 2% 4 5 2 3 1 6% 2% 38% 6% 48%
PLOT004 51% 25% 17% 6% 1% 1 3 4 2 5 51% 17% 6% 25% 1%
PLOT005 46% 33% 9% 9% 4% 5 1 3 4 2 4% 46% 9% 9% 33%
PLOT006 60% 26% 8% 5% 1% 1 4 3 5 2 60% 5% 8% 1% 26%
PLOT007 69% 13% 8% 6% 4% 3 5 2 4 1 8% 4% 13% 6% 69%
PLOT008 36% 22% 21% 12% 8% 3 1 5 4 2 21% 36% 8% 12% 22%
PLOT009 41% 30% 21% 6% 2% 4 1 5 3 2 6% 41% 2% 21% 30%
PLOT010 50% 27% 14% 7% 2% 1 4 3 2 5 50% 7% 14% 27% 2%

Number of cases read: 10 Number of cases listed: 10

===================
APPENDIX: TEST DATA
===================
* ............ Test data ............ .
* 05 species, 10 cases (plots) .
* Abundances exponentially distributed with mean 100, .
* normalized to total abundance 1 and converted to (UGH!) .
* SPSS 'percents' for display. .

NEW FILE.
INPUT PROGRAM.
. STRING PLOT_ID (A8).
. NUMERIC COVR01 TO COVR05 (F4).
. VECTOR COVER =
COVR01 TO COVR05.
. LOOP #PLOT = 1 TO 10.
. COMPUTE PLOT_ID = CONCAT('PLOT',STRING(#PLOT,N3)).
. LOOP #PLANT = 1 TO 5.
. COMPUTE COVER(#PLANT)=RV.EXP(.01).
. END LOOP.
* Normalize, to total cover = 1 . .
. COMPUTE #TOTAL = SUM(COVR01 TO COVR05).
. LOOP #PLANT = 1 TO 05.
* The multiplication by 100, below, is because of .
* SPSS's very unfortunate implementation of PCT .
* formats. Sigh. .
. COMPUTE COVER(#PLANT)=100*COVER(#PLANT)/#TOTAL.
. END LOOP.
. END CASE.
. END LOOP.
. FORMATS COVR01 TO COVR05 (PCT3).
END FILE.
END INPUT PROGRAM.

Peck, Jon

Re: Ranking within cases individually

accessType is a parameter new to SPSS 15, which added the ability to write new variables to the active file (accessType='w') or to append new cases (accessType='a'). In SPSS 14, you would need to do this indirectly by, for example, writing the cases to a text file and then reading that with a DATA LIST command.

I could post syntax for that if this is too mysterious and there is interest.

-Jon Peck

-----Original Message-----
From: Roberts, Michael [mailto:[hidden email]]
Sent: Monday, December 11, 2006 4:22 PM
To: Peck, Jon; [hidden email]
Subject: RE: Re: Ranking within cases individually

Jon,

What would I have to do to make this work in version 14.0.1? I tried
running the code but 'accessType' appears to generate a problem!

Thank You

Mike

-----Original Message-----
From: SPSSX(r) Discussion [mailto:[hidden email]] On Behalf Of
Peck, Jon
Sent: Monday, December 11, 2006 1:52 PM
To: [hidden email]
Subject: Re: Ranking within cases individually

Here is a Python solution assuming SPSS 15 and programmability are
installed.
First it defines a general function for this task and then it calls it
with two variables as an example.
The part in triple quotes is the documentation for the function. sysmis
values sort low.
It creates new variables with the sorted values.

The "for v in varlist" loop creates new variables; the "for case in
curs" loops over the cases

begin program.
import spss, spssdata

def vsort(varlist, direction='a', suffix='_sorted'):
"""sort variables, assumed to be numeric, within cases creating
new variables

varlist is the list of variables to sort.
direction can be 'a' or 'd' for ascending or descending.
suffix is the string to append to the original name for the
corresponding new variable."""

rev = direction =='d'
curs = spssdata.Spssdata(indexes=varlist, names=False,
accessType='w')
for v in varlist:
curs.append(v + suffix)
curs.commitdict()
for case in curs:
curs.casevalues(sorted(case, reverse = rev))
curs.Cclose()

vsort(['salbegin','salary'])
end program.

-Jon Peck

-----Original Message-----
From: SPSSX(r) Discussion [mailto:[hidden email]] On Behalf Of
Richard Ristow
Sent: Monday, December 11, 2006 12:14 PM
To: [hidden email]
Subject: Re: [SPSSX-L] Ranking within cases individually

(Marta and Jan, I'm copying to you because you're probably the best
MATRIX whizzes among us.)

At 03:00 PM 12/8/2006, Erin Miles wrote:

>I have a data set that gives me plant percent cover data for 2100 plots

>(rows) and there are about 250 potential species that can be present in

>each plot (rows). I need to sort/rank each row so that the percent
>covers go from largest to smallest.

For sorting variables within cases, the best technique that seems to
exist is to use the GRADE function in the MATRIX language, which
apparently calls an internal sort routine; this has been noted as the
Ristow-Peck algorithm. Below is an example, SPSS draft output. Only just
out of testing. There doesn't seem to be a way to do it without a LOOP
in MATRIX.

There's a lot of matrix tracing that could be omitted in production.

The MATRIX SAVE wipes the working file. To join to the pre-existing
file, save the latter and re-attach using MATCH FILES. And, here goes:

LIST.

Notes
|-----------------------------|---------------------------|
|Output Created |11-DEC-2006 13:03:27 |
|-----------------------------|---------------------------|
PLOT_ID COVR01 COVR02 COVR03 COVR04 COVR05

PLOT001 7% 26% 44% 16% 8%
PLOT002 51% 18% 25% 2% 5%
PLOT003 6% 2% 38% 6% 48%
PLOT004 51% 17% 6% 25% 1%
PLOT005 4% 46% 9% 9% 33%
PLOT006 60% 5% 8% 1% 26%
PLOT007 8% 4% 13% 6% 69%
PLOT008 21% 36% 8% 12% 22%
PLOT009 6% 41% 2% 21% 30%
PLOT010 50% 7% 14% 27% 2%

Number of cases read: 10 Number of cases listed: 10

MATRIX.
GET PLOT_ID /FILE=*/VARIABLES=PLOT_ID.
GET COVER /FILE=*/VARIABLES=COVR01 TO COVR05.
PRINT PLOT_ID /FORMAT='A8'.
PRINT COVER/FORMAT='F5.1'/RNAMES=PLOT_ID.

COMPUTE CVR_SORT = 0*COVER.
COMPUTE CVR_ORDR = 0*COVER.

LOOP ROW = 1 TO NROW(COVER).
PRINT ROW.
COMPUTE ROW_ORDR = GRADE(-1*COVER(ROW,:)).
COMPUTE CVR_ORDR(ROW,:) = Row_ORDR.
COMPUTE CVR_SORT(ROW,Row_Ordr) = COVER(Row,:).

COMPUTE ROW_RSLT ={COVER(ROW,:) ;
CVR_ORDR(ROW,:);
CVR_SORT(ROW,:)}.
PRINT ROW_RSLT /FORMAT=F5.1
/RLABELS='Input','Order','Sorted'.

END LOOP.

PRINT CVR_ORDR/RNAMES=PLOT_ID.
PRINT CVR_SORT/FORMAT='F5.1'/RNAMES=PLOT_ID.

SAVE {PLOT_ID,CVR_SORT,CVR_ORDR,COVER}
/OUTFILE=*
/VARIABLES=PLOT_ID,SORTED01 TO SORTED05,
ORDER01 TO ORDER05,
COVER01 To COVER05
/STRINGS=PLOT_ID.
END MATRIX.

Notes
|--------------------------|---------------------------|
|Output Created |11-DEC-2006 13:03:27 |
|--------------------------|---------------------------|
Run MATRIX procedure:

PLOT_ID
PLOT001
PLOT002
PLOT003
PLOT004
PLOT005
PLOT006
PLOT007
PLOT008
PLOT009
PLOT010

COVER
PLOT001 6.9 26.0 43.6 15.8 7.8
PLOT002 50.5 17.9 24.8 2.3 4.6
PLOT003 5.9 1.7 37.5 6.4 48.5
PLOT004 51.4 16.8 5.9 25.2 .7
PLOT005 3.7 45.9 8.9 8.7 32.8
PLOT006 59.7 5.2 8.0 1.1 26.0
PLOT007 7.8 3.7 12.8 6.4 69.2
PLOT008 21.4 36.5 7.8 12.4 21.9
PLOT009 6.4 41.0 1.9 20.7 30.0
PLOT010 50.4 7.0 13.6 26.6 2.5

ROW
1

ROW_RSLT
Input 6.9 26.0 43.6 15.8 7.8
Order 5.0 2.0 1.0 3.0 4.0
Sorted 43.6 26.0 15.8 7.8 6.9

ROW
2

ROW_RSLT
Input 50.5 17.9 24.8 2.3 4.6
Order 1.0 3.0 2.0 5.0 4.0
Sorted 50.5 24.8 17.9 4.6 2.3

ROW
3

ROW_RSLT
Input 5.9 1.7 37.5 6.4 48.5
Order 4.0 5.0 2.0 3.0 1.0
Sorted 48.5 37.5 6.4 5.9 1.7

ROW
4

ROW_RSLT
Input 51.4 16.8 5.9 25.2 .7
Order 1.0 3.0 4.0 2.0 5.0
Sorted 51.4 25.2 16.8 5.9 .7

ROW
5

ROW_RSLT
Input 3.7 45.9 8.9 8.7 32.8
Order 5.0 1.0 3.0 4.0 2.0
Sorted 45.9 32.8 8.9 8.7 3.7

ROW
6

ROW_RSLT
Input 59.7 5.2 8.0 1.1 26.0
Order 1.0 4.0 3.0 5.0 2.0
Sorted 59.7 26.0 8.0 5.2 1.1

ROW
7

ROW_RSLT
Input 7.8 3.7 12.8 6.4 69.2
Order 3.0 5.0 2.0 4.0 1.0
Sorted 69.2 12.8 7.8 6.4 3.7

ROW
8

ROW_RSLT
Input 21.4 36.5 7.8 12.4 21.9
Order 3.0 1.0 5.0 4.0 2.0
Sorted 36.5 21.9 21.4 12.4 7.8

ROW
9

ROW_RSLT
Input 6.4 41.0 1.9 20.7 30.0
Order 4.0 1.0 5.0 3.0 2.0
Sorted 41.0 30.0 20.7 6.4 1.9

ROW
10

ROW_RSLT
Input 50.4 7.0 13.6 26.6 2.5
Order 1.0 4.0 3.0 2.0 5.0
Sorted 50.4 26.6 13.6 7.0 2.5

CVR_ORDR
PLOT001 5 2 1 3 4
PLOT002 1 3 2 5 4
PLOT003 4 5 2 3 1
PLOT004 1 3 4 2 5
PLOT005 5 1 3 4 2
PLOT006 1 4 3 5 2
PLOT007 3 5 2 4 1
PLOT008 3 1 5 4 2
PLOT009 4 1 5 3 2
PLOT010 1 4 3 2 5

CVR_SORT
PLOT001 43.6 26.0 15.8 7.8 6.9
PLOT002 50.5 24.8 17.9 4.6 2.3
PLOT003 48.5 37.5 6.4 5.9 1.7
PLOT004 51.4 25.2 16.8 5.9 .7
PLOT005 45.9 32.8 8.9 8.7 3.7
PLOT006 59.7 26.0 8.0 5.2 1.1
PLOT007 69.2 12.8 7.8 6.4 3.7
PLOT008 36.5 21.9 21.4 12.4 7.8
PLOT009 41.0 30.0 20.7 6.4 1.9
PLOT010 50.4 26.6 13.6 7.0 2.5

------ END MATRIX -----

FORMATS ORDER01 TO ORDER05 (F3)
/SORTED01 TO SORTED05 (PCT3),
/COVER01 TO COVER05 (PCT3).
TEMPORARY.
STRING SPACE(A8).
LIST VARIABLES=PLOT_ID
SORTED01 TO SORTED05
ORDER01 TO ORDER05
COVER01 TO COVER05
SPACE.

Notes
|-----------------------------|---------------------------|
|Output Created |11-DEC-2006 13:03:30 |
|-----------------------------|---------------------------|
SOR SOR SOR SOR SOR ORD ORD ORD ORD ORD COV COV
TED TED TED TED TED ER0 ER0 ER0 ER0 ER0 ER0 ER0 COVE COVE COVE
PLOT_ID 01 02 03 04 05 1 2 3 4 5 1 2 R03 R04 R05
SPACE

PLOT001 44% 26% 16% 8% 7% 5 2 1 3 4 7% 26% 44% 16% 8%
PLOT002 51% 25% 18% 5% 2% 1 3 2 5 4 51% 18% 25% 2% 5%
PLOT003 48% 38% 6% 6% 2% 4 5 2 3 1 6% 2% 38% 6% 48%
PLOT004 51% 25% 17% 6% 1% 1 3 4 2 5 51% 17% 6% 25% 1%
PLOT005 46% 33% 9% 9% 4% 5 1 3 4 2 4% 46% 9% 9% 33%
PLOT006 60% 26% 8% 5% 1% 1 4 3 5 2 60% 5% 8% 1% 26%
PLOT007 69% 13% 8% 6% 4% 3 5 2 4 1 8% 4% 13% 6% 69%
PLOT008 36% 22% 21% 12% 8% 3 1 5 4 2 21% 36% 8% 12% 22%
PLOT009 41% 30% 21% 6% 2% 4 1 5 3 2 6% 41% 2% 21% 30%
PLOT010 50% 27% 14% 7% 2% 1 4 3 2 5 50% 7% 14% 27% 2%

Number of cases read: 10 Number of cases listed: 10

===================
APPENDIX: TEST DATA
===================
* ............ Test data ............ .
* 05 species, 10 cases (plots) .
* Abundances exponentially distributed with mean 100, .
* normalized to total abundance 1 and converted to (UGH!) .
* SPSS 'percents' for display. .

NEW FILE.
INPUT PROGRAM.
. STRING PLOT_ID (A8).
. NUMERIC COVR01 TO COVR05 (F4).
. VECTOR COVER =
COVR01 TO COVR05.
. LOOP #PLOT = 1 TO 10.
. COMPUTE PLOT_ID = CONCAT('PLOT',STRING(#PLOT,N3)).
. LOOP #PLANT = 1 TO 5.
. COMPUTE COVER(#PLANT)=RV.EXP(.01).
. END LOOP.
* Normalize, to total cover = 1 . .
. COMPUTE #TOTAL = SUM(COVR01 TO COVR05).
. LOOP #PLANT = 1 TO 05.
* The multiplication by 100, below, is because of .
* SPSS's very unfortunate implementation of PCT .
* formats. Sigh. .
. COMPUTE COVER(#PLANT)=100*COVER(#PLANT)/#TOTAL.
. END LOOP.
. END CASE.
. END LOOP.
. FORMATS COVR01 TO COVR05 (PCT3).
END FILE.
END INPUT PROGRAM.

Gary Oliver

Graphing Multiple Variables using the one command

In reply to this post by Roberts, Michael

Colleagues

I am checking some of the variables in my correlation data to get a
graph of the percentage of responses at each of the seven items on a
Likert scale. I am using the following command:
GRAPH
/BAR(SIMPLE)=PCT BY <NAME OF VARIABLE> .
In this syntax I have to run the command for each variable separately.
Is there a way of including all the variable names in the one command or
must I have a separate command syntax for each variable.

Warm regards/gary

Unit Co-ordinator for Business Information Systems
(A Postgraduate Master of Commerce and Master of Business unit)
School of Business
The University of Sydney
------------------------
,-_|\ Building H69, Office 437
/ \ Corner Codrington Street
\_,-._* & Rose Street
Darlington 2006
@ Australia
--------------------------------------
E-mail: [hidden email]
------------------------
Location details:
Travelling from Broadway, turn south off City Road
Navigate toward the Acquatic Centre
------------------------
University Map: http://db.auth.usyd.edu.au/directories/map/index.stm
University Website:
www.usyd.edu.au
Faculty Website
www.econ.usyd.edu.au
------------------------
Faculty Student Information Office
(Timetables, Special Consideration)
Merewether Building
Enter from City Road side
e-mail: [hidden email]
Phone: 9351-3076
----------------------------------
Executive Officer for Business Information Systems
Katy Roy
Room 347, Building H69
E-mail: [hidden email]
Phone: 9036 9432
---------------------------------

>>-----Original Message-----
>>From: SPSSX(r) Discussion [mailto:[hidden email]]
>>On Behalf Of Roberts, Michael
>>Sent: Tuesday, December 12, 2006 9:22 AM
>>To: [hidden email]
>>Subject: Re: Ranking within cases individually
>>
>> Jon,
>>
>>What would I have to do to make this work in version 14.0.1?
>>I tried running the code but 'accessType' appears to generate
>>a problem!
>>
>>Thank You
>>
>>Mike
>>
>>
>>-----Original Message-----
>>From: SPSSX(r) Discussion [mailto:[hidden email]]
>>On Behalf Of Peck, Jon
>>Sent: Monday, December 11, 2006 1:52 PM
>>To: [hidden email]
>>Subject: Re: Ranking within cases individually
>>
>>Here is a Python solution assuming SPSS 15 and
>>programmability are installed.
>>First it defines a general function for this task and then it
>>calls it with two variables as an example.
>>The part in triple quotes is the documentation for the
>>function. sysmis values sort low.
>>It creates new variables with the sorted values.
>>
>>The "for v in varlist" loop creates new variables; the "for
>>case in curs" loops over the cases
>>
>>begin program.
>>import spss, spssdata
>>
>>def vsort(varlist, direction='a', suffix='_sorted'):
>> """sort variables, assumed to be numeric, within
>>cases creating new variables
>>
>> varlist is the list of variables to sort.
>> direction can be 'a' or 'd' for ascending or descending.
>> suffix is the string to append to the original name for the
>> corresponding new variable."""
>>
>> rev = direction =='d'
>> curs = spssdata.Spssdata(indexes=varlist, names=False,
>>accessType='w')
>> for v in varlist:
>> curs.append(v + suffix)
>> curs.commitdict()
>> for case in curs:
>> curs.casevalues(sorted(case, reverse = rev))
>> curs.Cclose()
>>
>>vsort(['salbegin','salary'])
>>end program.
>>
>>-Jon Peck
>>
>>-----Original Message-----
>>From: SPSSX(r) Discussion [mailto:[hidden email]]
>>On Behalf Of Richard Ristow
>>Sent: Monday, December 11, 2006 12:14 PM
>>To: [hidden email]
>>Subject: Re: [SPSSX-L] Ranking within cases individually
>>
>>(Marta and Jan, I'm copying to you because you're probably
>>the best MATRIX whizzes among us.)
>>
>>At 03:00 PM 12/8/2006, Erin Miles wrote:
>>
>>>I have a data set that gives me plant percent cover data for
>>2100 plots
>>
>>>(rows) and there are about 250 potential species that can be
>>present in
>>
>>>each plot (rows). I need to sort/rank each row so that the percent
>>>covers go from largest to smallest.
>>
>>For sorting variables within cases, the best technique that
>>seems to exist is to use the GRADE function in the MATRIX
>>language, which apparently calls an internal sort routine;
>>this has been noted as the Ristow-Peck algorithm. Below is an
>>example, SPSS draft output. Only just out of testing. There
>>doesn't seem to be a way to do it without a LOOP in MATRIX.
>>
>>There's a lot of matrix tracing that could be omitted in production.
>>
>>The MATRIX SAVE wipes the working file. To join to the
>>pre-existing file, save the latter and re-attach using MATCH
>>FILES. And, here goes:
>>
>>LIST.
>>
>>Notes
>>|-----------------------------|---------------------------|
>>|Output Created |11-DEC-2006 13:03:27 |
>>|-----------------------------|---------------------------|
>>PLOT_ID COVR01 COVR02 COVR03 COVR04 COVR05
>>
>>PLOT001 7% 26% 44% 16% 8%
>>PLOT002 51% 18% 25% 2% 5%
>>PLOT003 6% 2% 38% 6% 48%
>>PLOT004 51% 17% 6% 25% 1%
>>PLOT005 4% 46% 9% 9% 33%
>>PLOT006 60% 5% 8% 1% 26%
>>PLOT007 8% 4% 13% 6% 69%
>>PLOT008 21% 36% 8% 12% 22%
>>PLOT009 6% 41% 2% 21% 30%
>>PLOT010 50% 7% 14% 27% 2%
>>
>>Number of cases read: 10 Number of cases listed: 10
>>
>>
>>MATRIX.
>>GET PLOT_ID /FILE=*/VARIABLES=PLOT_ID.
>>GET COVER /FILE=*/VARIABLES=COVR01 TO COVR05.
>>PRINT PLOT_ID /FORMAT='A8'.
>>PRINT COVER/FORMAT='F5.1'/RNAMES=PLOT_ID.
>>
>>COMPUTE CVR_SORT = 0*COVER.
>>COMPUTE CVR_ORDR = 0*COVER.
>>
>>LOOP ROW = 1 TO NROW(COVER).
>>PRINT ROW.
>>COMPUTE ROW_ORDR = GRADE(-1*COVER(ROW,:)).
>>COMPUTE CVR_ORDR(ROW,:) = Row_ORDR.
>>COMPUTE CVR_SORT(ROW,Row_Ordr) = COVER(Row,:).
>>
>>
>>COMPUTE ROW_RSLT ={COVER(ROW,:) ;
>> CVR_ORDR(ROW,:);
>> CVR_SORT(ROW,:)}.
>>PRINT ROW_RSLT /FORMAT=F5.1
>> /RLABELS='Input','Order','Sorted'.
>>
>>END LOOP.
>>
>>PRINT CVR_ORDR/RNAMES=PLOT_ID.
>>PRINT CVR_SORT/FORMAT='F5.1'/RNAMES=PLOT_ID.
>>
>>SAVE {PLOT_ID,CVR_SORT,CVR_ORDR,COVER}
>> /OUTFILE=*
>> /VARIABLES=PLOT_ID,SORTED01 TO SORTED05,
>> ORDER01 TO ORDER05,
>> COVER01 To COVER05
>> /STRINGS=PLOT_ID.
>>END MATRIX.
>>
>>Notes
>>|--------------------------|---------------------------|
>>|Output Created |11-DEC-2006 13:03:27 |
>>|--------------------------|---------------------------|
>>Run MATRIX procedure:
>>
>>PLOT_ID
>> PLOT001
>> PLOT002
>> PLOT003
>> PLOT004
>> PLOT005
>> PLOT006
>> PLOT007
>> PLOT008
>> PLOT009
>> PLOT010
>>
>>COVER
>>PLOT001 6.9 26.0 43.6 15.8 7.8
>>PLOT002 50.5 17.9 24.8 2.3 4.6
>>PLOT003 5.9 1.7 37.5 6.4 48.5
>>PLOT004 51.4 16.8 5.9 25.2 .7
>>PLOT005 3.7 45.9 8.9 8.7 32.8
>>PLOT006 59.7 5.2 8.0 1.1 26.0
>>PLOT007 7.8 3.7 12.8 6.4 69.2
>>PLOT008 21.4 36.5 7.8 12.4 21.9
>>PLOT009 6.4 41.0 1.9 20.7 30.0
>>PLOT010 50.4 7.0 13.6 26.6 2.5
>>
>>ROW
>> 1
>>
>>ROW_RSLT
>>Input 6.9 26.0 43.6 15.8 7.8
>>Order 5.0 2.0 1.0 3.0 4.0
>>Sorted 43.6 26.0 15.8 7.8 6.9
>>
>>ROW
>> 2
>>
>>ROW_RSLT
>>Input 50.5 17.9 24.8 2.3 4.6
>>Order 1.0 3.0 2.0 5.0 4.0
>>Sorted 50.5 24.8 17.9 4.6 2.3
>>
>>ROW
>> 3
>>
>>ROW_RSLT
>>Input 5.9 1.7 37.5 6.4 48.5
>>Order 4.0 5.0 2.0 3.0 1.0
>>Sorted 48.5 37.5 6.4 5.9 1.7
>>
>>ROW
>> 4
>>
>>ROW_RSLT
>>Input 51.4 16.8 5.9 25.2 .7
>>Order 1.0 3.0 4.0 2.0 5.0
>>Sorted 51.4 25.2 16.8 5.9 .7
>>
>>ROW
>> 5
>>
>>ROW_RSLT
>>Input 3.7 45.9 8.9 8.7 32.8
>>Order 5.0 1.0 3.0 4.0 2.0
>>Sorted 45.9 32.8 8.9 8.7 3.7
>>
>>ROW
>> 6
>>
>>ROW_RSLT
>>Input 59.7 5.2 8.0 1.1 26.0
>>Order 1.0 4.0 3.0 5.0 2.0
>>Sorted 59.7 26.0 8.0 5.2 1.1
>>
>>ROW
>> 7
>>
>>ROW_RSLT
>>Input 7.8 3.7 12.8 6.4 69.2
>>Order 3.0 5.0 2.0 4.0 1.0
>>Sorted 69.2 12.8 7.8 6.4 3.7
>>
>>ROW
>> 8
>>
>>ROW_RSLT
>>Input 21.4 36.5 7.8 12.4 21.9
>>Order 3.0 1.0 5.0 4.0 2.0
>>Sorted 36.5 21.9 21.4 12.4 7.8
>>
>>ROW
>> 9
>>
>>ROW_RSLT
>>Input 6.4 41.0 1.9 20.7 30.0
>>Order 4.0 1.0 5.0 3.0 2.0
>>Sorted 41.0 30.0 20.7 6.4 1.9
>>
>>ROW
>> 10
>>
>>ROW_RSLT
>>Input 50.4 7.0 13.6 26.6 2.5
>>Order 1.0 4.0 3.0 2.0 5.0
>>Sorted 50.4 26.6 13.6 7.0 2.5
>>
>>CVR_ORDR
>>PLOT001 5 2 1 3 4
>>PLOT002 1 3 2 5 4
>>PLOT003 4 5 2 3 1
>>PLOT004 1 3 4 2 5
>>PLOT005 5 1 3 4 2
>>PLOT006 1 4 3 5 2
>>PLOT007 3 5 2 4 1
>>PLOT008 3 1 5 4 2
>>PLOT009 4 1 5 3 2
>>PLOT010 1 4 3 2 5
>>
>>CVR_SORT
>>PLOT001 43.6 26.0 15.8 7.8 6.9
>>PLOT002 50.5 24.8 17.9 4.6 2.3
>>PLOT003 48.5 37.5 6.4 5.9 1.7
>>PLOT004 51.4 25.2 16.8 5.9 .7
>>PLOT005 45.9 32.8 8.9 8.7 3.7
>>PLOT006 59.7 26.0 8.0 5.2 1.1
>>PLOT007 69.2 12.8 7.8 6.4 3.7
>>PLOT008 36.5 21.9 21.4 12.4 7.8
>>PLOT009 41.0 30.0 20.7 6.4 1.9
>>PLOT010 50.4 26.6 13.6 7.0 2.5
>>
>>------ END MATRIX -----
>>
>>
>>FORMATS ORDER01 TO ORDER05 (F3)
>> /SORTED01 TO SORTED05 (PCT3),
>> /COVER01 TO COVER05 (PCT3).
>>TEMPORARY.
>>STRING SPACE(A8).
>>LIST VARIABLES=PLOT_ID
>> SORTED01 TO SORTED05
>> ORDER01 TO ORDER05
>> COVER01 TO COVER05
>> SPACE.
>>
>>Notes
>>|-----------------------------|---------------------------|
>>|Output Created |11-DEC-2006 13:03:30 |
>>|-----------------------------|---------------------------|
>> SOR SOR SOR SOR SOR ORD ORD ORD ORD ORD COV COV
>> TED TED TED TED TED ER0 ER0 ER0 ER0 ER0 ER0 ER0
>>COVE COVE COVE
>>PLOT_ID 01 02 03 04 05 1 2 3 4 5 1 2 R03 R04 R05
>>SPACE
>>
>>PLOT001 44% 26% 16% 8% 7% 5 2 1 3 4 7% 26% 44%
>> 16% 8%
>>PLOT002 51% 25% 18% 5% 2% 1 3 2 5 4 51% 18% 25%
>> 2% 5%
>>PLOT003 48% 38% 6% 6% 2% 4 5 2 3 1 6% 2% 38%
>> 6% 48%
>>PLOT004 51% 25% 17% 6% 1% 1 3 4 2 5 51% 17% 6%
>> 25% 1%
>>PLOT005 46% 33% 9% 9% 4% 5 1 3 4 2 4% 46% 9%
>> 9% 33%
>>PLOT006 60% 26% 8% 5% 1% 1 4 3 5 2 60% 5% 8%
>> 1% 26%
>>PLOT007 69% 13% 8% 6% 4% 3 5 2 4 1 8% 4% 13%
>> 6% 69%
>>PLOT008 36% 22% 21% 12% 8% 3 1 5 4 2 21% 36% 8%
>> 12% 22%
>>PLOT009 41% 30% 21% 6% 2% 4 1 5 3 2 6% 41% 2%
>> 21% 30%
>>PLOT010 50% 27% 14% 7% 2% 1 4 3 2 5 50% 7% 14%
>> 27% 2%
>>
>>Number of cases read: 10 Number of cases listed: 10
>>
>>===================
>>APPENDIX: TEST DATA
>>===================
>>* ............ Test data ............ .
>>* 05 species, 10 cases (plots) .
>>* Abundances exponentially distributed with mean 100, .
>>* normalized to total abundance 1 and converted to (UGH!) .
>>* SPSS 'percents' for display. .
>>
>>NEW FILE.
>>INPUT PROGRAM.
>>. STRING PLOT_ID (A8).
>>. NUMERIC COVR01 TO COVR05 (F4).
>>. VECTOR COVER =
>> COVR01 TO COVR05.
>>. LOOP #PLOT = 1 TO 10.
>>. COMPUTE PLOT_ID = CONCAT('PLOT',STRING(#PLOT,N3)).
>>. LOOP #PLANT = 1 TO 5.
>>. COMPUTE COVER(#PLANT)=RV.EXP(.01).
>>. END LOOP.
>>* Normalize, to total cover = 1 . .
>>. COMPUTE #TOTAL = SUM(COVR01 TO COVR05).
>>. LOOP #PLANT = 1 TO 05.
>>* The multiplication by 100, below, is because of .
>>* SPSS's very unfortunate implementation of PCT .
>>* formats. Sigh. .
>>. COMPUTE COVER(#PLANT)=100*COVER(#PLANT)/#TOTAL.
>>. END LOOP.
>>. END CASE.
>>. END LOOP.
>>. FORMATS COVR01 TO COVR05 (PCT3).
>>END FILE.
>>END INPUT PROGRAM.
>>

Gary Oliver

Reporting on all Values for Variables

In reply to this post by Roberts, Michael

Colleagues

In my last inquiry I was trying to make the simple bar graph syntax work
for more than one variable.

Having looked at the output for both the GRAPH and the FREQUENCIES
command, it seems to me that SPSS will only report for each of the scale
points within my Likert if there is a case using that value. That is, it
will report: extremely undesirable=101, neutral=183 but not:
undesirable=0, desirable=0, extremely desirable=0.

My question is: Is it possible to force SPSS to report where the
percentage (or count) is none?

Warm regards/gary

Unit Co-ordinator for Business Information Systems
(A Postgraduate Master of Commerce and Master of Business unit)
School of Business
The University of Sydney
------------------------
,-_|\ Building H69, Office 437
/ \ Corner Codrington Street
\_,-._* & Rose Street
Darlington 2006
@ Australia
--------------------------------------
E-mail: [hidden email]
------------------------
Location details:
Travelling from Broadway, turn south off City Road
Navigate toward the Acquatic Centre
------------------------
University Map: http://db.auth.usyd.edu.au/directories/map/index.stm
University Website:
www.usyd.edu.au
Faculty Website
www.econ.usyd.edu.au
------------------------
Faculty Student Information Office
(Timetables, Special Consideration)
Merewether Building
Enter from City Road side
e-mail: [hidden email]
Phone: 9351-3076
----------------------------------
Executive Officer for Business Information Systems
Katy Roy
Room 347, Building H69
E-mail: [hidden email]
Phone: 9036 9432
---------------------------------

Marta García-Granero

Re: Reporting on all Values for Variables

Hi Gary

Wednesday, December 13, 2006, 7:39:22 AM, You wrote:

GO> In my last inquiry I was trying to make the simple bar graph syntax work
GO> for more than one variable.

GO> Having looked at the output for both the GRAPH and the FREQUENCIES
GO> command, it seems to me that SPSS will only report for each of the scale
GO> points within my Likert if there is a case using that value. That is, it
GO> will report: extremely undesirable=101, neutral=183 but not:
GO> undesirable=0, desirable=0, extremely desirable=0.

GO> My question is: Is it possible to force SPSS to report where the
GO> percentage (or count) is none?

You can to your file 5 cases corresponding to each of the categories,
give them a very low weight and then runn the frequencies/graph you
want. The following code solves it assuming that your likert scale
goes from 1 to 5:

* Fake data (replace by your own) *.
DATA LIST LIST/var1 TO var7 (7 F8).
BEGIN DATA
1 2 4 1 5 1 3
3 5 3 5 4 4 2
END DATA.

VALUE LABEL var1 TO var4
1 'Extremely undesirable'
2 'Undesirable'
3 'Neutral'
4 'Desirable'
5 'Extremely desirable'.

FREQ VAR=ALL.

MATRIX.
GET data /VAR=ALL /NAMES=vnames /MISSING=OMIT.
COMPUTE k=NCOL(data).
COMPUTE Emptycat=T(MAKE(k,1,1)*{1:5}).
COMPUTE weight=MAKE(5,1,0.000001).
Compute AllNames={vnames,'weight'}.
SAVE {EmptyCat,weight} /OUTFILE='C:\Temp\EmptyCategories.sav' /NAMES=AllNames.
END MATRIX.

COMPUTE WEIGHT=1.
ADD FILES /FILE=*
/FILE='C:\temp\EmptyCategories.sav'.
WEIGHT BY weight.

* See result *.
FREQ VAR=ALL.

GRAPH /BAR(SIMPLE)=PCT BY var1.

Regards,
Marta

John Painter

survival analysis question

In reply to this post by Peck, Jon

Dear List,

Please pardon my posing this question here, but apparently the answer is so
obvious (to everyone but me) that it is not worthy of discussion in any of
the survival analysis material I have at hand.

Q: For cases that are (right) censored is the time variable set to missing,
or to date of data collection / end of study?

Thanks in advance,
John

Melissa Ives

Re: survival analysis question

Set it to the last chance of knowing whether the terminal event occurred
(e.g. so you can say that as of MM/DD/YYYY, the terminal event had NOT
occurred.) I am not sure if that is the end of the study or the last
date of data collection for the client?

Melissa

-----Original Message-----
From: SPSSX(r) Discussion [mailto:[hidden email]] On Behalf Of
John Painter
Sent: Thursday, January 04, 2007 2:47 PM
To: [hidden email]
Subject: [SPSSX-L] survival analysis question

Dear List,

Please pardon my posing this question here, but apparently the answer is
so obvious (to everyone but me) that it is not worthy of discussion in
any of the survival analysis material I have at hand.

Q: For cases that are (right) censored is the time variable set to
missing, or to date of data collection / end of study?

Thanks in advance,
John

PRIVILEGED AND CONFIDENTIAL INFORMATION
This transmittal and any attachments may contain PRIVILEGED AND
CONFIDENTIAL information and is intended only for the use of the
addressee. If you are not the designated recipient, or an employee
or agent authorized to deliver such transmittals to the designated
recipient, you are hereby notified that any dissemination,
copying or publication of this transmittal is strictly prohibited. If
you have received this transmittal in error, please notify us
immediately by replying to the sender and delete this copy from your
system. You may also call us at (309) 827-6026 for assistance.

Hector Maletta

Re: survival analysis question

In reply to this post by John Painter

The time variable should be set at the date of end of study, and the
event variable to zero. This means that at the end of the study the subject
had not yet undergone the event (e.g. he/she was still alive). If the time
variable is missing, the case would be excluded from the analysis, instead
of been treated as censored.

Hector

-----Mensaje original-----
De: SPSSX(r) Discussion [mailto:[hidden email]] En nombre de John
Painter
Enviado el: 04 January 2007 17:47
Para: [hidden email]
Asunto: survival analysis question

Dear List,

Please pardon my posing this question here, but apparently the
answer is so
obvious (to everyone but me) that it is not worthy of discussion in
any of
the survival analysis material I have at hand.

Q: For cases that are (right) censored is the time variable set to
missing,
or to date of data collection / end of study?

Thanks in advance,
John

Hector Maletta

Re: survival analysis question

In reply to this post by Melissa Ives

Melissa is right. In my answer I implicitly assumed continuous
observation, so the last time of observation was also the end of the study,
but for discrete observations the last time a subject was observed or
checked may have been before the official end of the study.
Hector

-----Mensaje original-----
De: SPSSX(r) Discussion [mailto:[hidden email]] En nombre de
Melissa Ives
Enviado el: 04 January 2007 17:53
Para: [hidden email]
Asunto: Re: survival analysis question

Set it to the last chance of knowing whether the terminal event
occurred
(e.g. so you can say that as of MM/DD/YYYY, the terminal event had
NOT
occurred.) I am not sure if that is the end of the study or the
last
date of data collection for the client?

Melissa

-----Original Message-----
From: SPSSX(r) Discussion [mailto:[hidden email]] On
Behalf Of
John Painter
Sent: Thursday, January 04, 2007 2:47 PM
To: [hidden email]
Subject: [SPSSX-L] survival analysis question

Dear List,

Please pardon my posing this question here, but apparently the
answer is
so obvious (to everyone but me) that it is not worthy of discussion
in
any of the survival analysis material I have at hand.

Q: For cases that are (right) censored is the time variable set to
missing, or to date of data collection / end of study?

Thanks in advance,
John

PRIVILEGED AND CONFIDENTIAL INFORMATION
This transmittal and any attachments may contain PRIVILEGED AND
CONFIDENTIAL information and is intended only for the use of the
addressee. If you are not the designated recipient, or an employee
or agent authorized to deliver such transmittals to the designated
recipient, you are hereby notified that any dissemination,
copying or publication of this transmittal is strictly prohibited.
If
you have received this transmittal in error, please notify us
immediately by replying to the sender and delete this copy from your
system. You may also call us at (309) 827-6026 for assistance.