SPSSX Discussion

David's method of generating a "counter" variable

Classic

List

Threaded

3 messages Options

Bruce Weaver

David's method of generating a "counter" variable

Administrator

I thought David's method of generating a "counter" variable was rather neat. The method I typically use (MATCH FILES followed by an IF statement) is a two-liner, whereas David's is a two-liner. Here's David's method with an extension for the case where one wants a counter within each of the V1*V2 cells (or combinations).

* David Marso's method for generating a counter variable.
* Posted to SPSSX-L mailing list on 30-Dec-2012 in thread "recode variable".

* Simulate some data.
new file.
- input program.
- loop ID = 1 to 1000.
- compute v1 = TRUNC(UNIFORM(4))+1.
- compute v2 = TRUNC(UNIFORM(4))+1.
- end case.
- end loop.
end file.
end input program.
execute.

********************************************************* .
* Count cases within levels of V1, ignoring V2 .
sort cases by v1.
* David's one-liner.
COMPUTE DMcount1=SUM(1,LAG(DMcount1)*(v1 EQ LAG(v1))).
* My customary two-liner.
match files file = * / by v1 / first = BWcount1 .
if (BWcount1 EQ 0) BWcount1 = LAG(BWcount1) + 1.
********************************************************* .
* Count cases within each V1*V2 cell .
sort cases by v1 v2.
* David's one-liner.
COMPUTE DMcount2=SUM(1,LAG(DMcount2)*(v1 EQ LAG(v1) AND v2 EQ LAG(v2))).
* My customary two-liner.
match files file = * / by v1 v2 / first = BWcount2 .
if (BWcount2 EQ 0) BWcount2 = LAG(BWcount2) + 1.
********************************************************* .

formats ID to BWcount2 (f5.0).

* Display maximum value of counter1 for each level of V1,
and verify results via FREQUENCIES.
means DMcount1 BWcount1 by v1 / cells = max.
frequencies v1.

* Display maximum value of counter2 for each V1*V2 cell,
and verify results via CROSSTABS.
means DMcount2 BWcount2 by v1 by v2 / cells = max.
CROSSTABS v1 by v2.

--
Bruce Weaver
bweaver@lakeheadu.ca
http://sites.google.com/a/lakeheadu.ca/bweaver/

"When all else fails, RTFM."

PLEASE NOTE THE FOLLOWING:
1. My Hotmail account is not monitored regularly. To send me an e-mail, please use the address shown above.
2. The SPSSX Discussion forum on Nabble is no longer linked to the SPSSX-L listserv administered by UGA (https://listserv.uga.edu/).

David Marso

Re: David's method of generating a "counter" variable

Administrator

There are a few other techniques which come to mind.
I post 6 here which include Bruce's and my versions previously posted.
new file.
- input program.
- loop ID = 1 to 1000.
- compute v1 = TRUNC(UNIFORM(4))+1.
- compute v2 = TRUNC(UNIFORM(4))+1.
- end case.
- end loop.
end file.
end input program.
****************************.
** 1 **.
COMPUTE XX=UNIFORM(1).
RANK XX BY V1 V2 /RANK INTO Counter1.
SORT CASES BY V1 V2 Counter1.

** 2 **.
SORT CASES BY V1 V2.
SPLIT FILE BY V1 V2.
COMPUTE XX=1.
CREATE COUNTER2=CSUM(XX).

** 3 **.
SORT CASES BY V1 V2.
MATCH FILES / FILE * / BY V1 V2 / FIRST=COUNTER3.
IF COUNTER3=0 COUNTER3=LAG(COUNTER3) + 1.

** 4 **.
SORT CASES BY V1 V2.
COMPUTE Counter4=SUM(1,LAG(Counter4)*(V1 EQ LAG(V1) AND V2 EQ LAG(V2))).

** 5 **.
SORT CASES BY V1 V2.
IF $CASENUM EQ 1 OR V1 NE LAG(V1) OR V2 NE LAG(V2) Counter5=1.
IF MISSING(Counter5) Counter5=LAG(Counter5)+1.

** 6 **.
SORT CASES BY V1 V2.
IF ANY(1,$CASENUM,V1 NE LAG(V1),V2 NE LAG(V2)) Counter6=1.
IF MISSING(Counter6) Counter6=LAG(Counter6)+1.

*******.
EXE.

Bruce Weaver wrote

I thought David's method of generating a "counter" variable was rather neat. The method I typically use (MATCH FILES followed by an IF statement) is a two-liner, whereas David's is a two-liner. Here's David's method with an extension for the case where one wants a counter within each of the V1*V2 cells (or combinations).

* David Marso's method for generating a counter variable.
* Posted to SPSSX-L mailing list on 30-Dec-2012 in thread "recode variable".

* Simulate some data.
new file.
- input program.
- loop ID = 1 to 1000.
- compute v1 = TRUNC(UNIFORM(4))+1.
- compute v2 = TRUNC(UNIFORM(4))+1.
- end case.
- end loop.
end file.
end input program.
execute.

********************************************************* .
* Count cases within levels of V1, ignoring V2 .
sort cases by v1.
* David's one-liner.
COMPUTE DMcount1=SUM(1,LAG(DMcount1)*(v1 EQ LAG(v1))).
* My customary two-liner.
match files file = * / by v1 / first = BWcount1 .
if (BWcount1 EQ 0) BWcount1 = LAG(BWcount1) + 1.
********************************************************* .
* Count cases within each V1*V2 cell .
sort cases by v1 v2.
* David's one-liner.
COMPUTE DMcount2=SUM(1,LAG(DMcount2)*(v1 EQ LAG(v1) AND v2 EQ LAG(v2))).
* My customary two-liner.
match files file = * / by v1 v2 / first = BWcount2 .
if (BWcount2 EQ 0) BWcount2 = LAG(BWcount2) + 1.
********************************************************* .

formats ID to BWcount2 (f5.0).

* Display maximum value of counter1 for each level of V1,
and verify results via FREQUENCIES.
means DMcount1 BWcount1 by v1 / cells = max.
frequencies v1.

* Display maximum value of counter2 for each V1*V2 cell,
and verify results via CROSSTABS.
means DMcount2 BWcount2 by v1 by v2 / cells = max.
CROSSTABS v1 by v2.

Please reply to the list and not to my personal email.
Those desiring my consulting or training services please feel free to email me.
---
"Nolite dare sanctum canibus neque mittatis margaritas vestras ante porcos ne forte conculcent eas pedibus suis."
Cum es damnatorum possederunt porcos iens ut salire off sanguinum cliff in abyssum?"

Tian Qiu

Re: David's method of generating a "counter" variable

Many thanks to both of you. These are really helpful.

Regards,
TQ

-----Original Message-----
From: SPSSX(r) Discussion [mailto:[hidden email]] On Behalf Of David Marso
Sent: 31 January 2012 14:13
To: [hidden email]
Subject: Re: David's method of generating a "counter" variable

There are a few other techniques which come to mind.
I post 6 here which include Bruce's and my versions previously posted.
new file.
- input program.
- loop ID = 1 to 1000.
- compute v1 = TRUNC(UNIFORM(4))+1.
- compute v2 = TRUNC(UNIFORM(4))+1.
- end case.
- end loop.
end file.
end input program.
****************************.
** 1 **.
COMPUTE XX=UNIFORM(1).
RANK XX BY V1 V2 /RANK INTO Counter1.
SORT CASES BY V1 V2 Counter1.

** 2 **.
SORT CASES BY V1 V2.
SPLIT FILE BY V1 V2.
COMPUTE XX=1.
CREATE COUNTER2=CSUM(XX).

** 3 **.
SORT CASES BY V1 V2.
MATCH FILES / FILE * / BY V1 V2 / FIRST=COUNTER3.
IF COUNTER3=0 COUNTER3=LAG(COUNTER3) + 1.

** 4 **.
SORT CASES BY V1 V2.
COMPUTE Counter4=SUM(1,LAG(Counter4)*(V1 EQ LAG(V1) AND V2 EQ LAG(V2))).

** 5 **.
SORT CASES BY V1 V2.
IF $CASENUM EQ 1 OR V1 NE LAG(V1) OR V2 NE LAG(V2) Counter5=1.
IF MISSING(Counter5) Counter5=LAG(Counter5)+1.

** 6 **.
SORT CASES BY V1 V2.
IF ANY(1,$CASENUM,V1 NE LAG(V1),V2 NE LAG(V2)) Counter6=1.
IF MISSING(Counter6) Counter6=LAG(Counter6)+1.

*******.
EXE.

Bruce Weaver wrote

>
> I thought David's method of generating a "counter" variable was rather
> neat. The method I typically use (MATCH FILES followed by an IF
> statement) is a two-liner, whereas David's is a two-liner. Here's
> David's method with an extension for the case where one wants a
> counter within each of the V1*V2 cells (or combinations).
>
>
> * David Marso's method for generating a counter variable.
> * Posted to SPSSX-L mailing list on 30-Dec-2012 in thread "recode
> variable".
>
> * Simulate some data.
> new file.
> - input program.
> - loop ID = 1 to 1000.
> - compute v1 = TRUNC(UNIFORM(4))+1.
> - compute v2 = TRUNC(UNIFORM(4))+1.
> - end case.
> - end loop.
> end file.
> end input program.
> execute.
>
> ********************************************************* .
> * Count cases within levels of V1, ignoring V2 .
> sort cases by v1.
> * David's one-liner.
> COMPUTE DMcount1=SUM(1,LAG(DMcount1)*(v1 EQ LAG(v1))).
> * My customary two-liner.
> match files file = * / by v1 / first = BWcount1 .
> if (BWcount1 EQ 0) BWcount1 = LAG(BWcount1) + 1.
> ********************************************************* .
> * Count cases within each V1*V2 cell .
> sort cases by v1 v2.
> * David's one-liner.
> COMPUTE DMcount2=SUM(1,LAG(DMcount2)*(v1 EQ LAG(v1) AND v2 EQ LAG(v2))).
> * My customary two-liner.
> match files file = * / by v1 v2 / first = BWcount2 .
> if (BWcount2 EQ 0) BWcount2 = LAG(BWcount2) + 1.
> ********************************************************* .
>
> formats ID to BWcount2 (f5.0).
>
> * Display maximum value of counter1 for each level of V1,
> and verify results via FREQUENCIES.
> means DMcount1 BWcount1 by v1 / cells = max.
> frequencies v1.
>
> * Display maximum value of counter2 for each V1*V2 cell,
> and verify results via CROSSTABS.
> means DMcount2 BWcount2 by v1 by v2 / cells = max.
> CROSSTABS v1 by v2.
>

--
View this message in context: http://spssx-discussion.1045642.n5.nabble.com/David-s-method-of-generating-a-counter-variable-tp5442896p5444626.html
Sent from the SPSSX Discussion mailing list archive at Nabble.com.

=====================
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

=====================
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