SPSSX Discussion

Recode with part of data in variable name?

Classic

List

Threaded

6 messages Options

lts1

Recode with part of data in variable name?

Hi SPSSers

I have a dataset with over 1000 cases each containing a four digit numeric code ranging from 111 to 9711. I need to recode these into a series of dummy variables based on the first 2 digits. I would like to combine the name of the code with the digit numbers. For example 111 would become Code01 and 9711 would become Code97. I started doing this as a series of IF statements, but it's long & tedious. I know there has to be an easier way. Unfortunately I don't know how to move the digits into the variable name. Can anyone help me? Thanks in advance.

Best,
Lisa

Lisa T. Stickney
Ph.D. Candidate
The Fox School of Business
and Management
Temple University
[hidden email]

Anton Balabanov

Re: Recode with part of data in variable name?

Hi, Lisa,

This syntax is far from general, but goes in your case, is it?
If your codes are sparse, it produces too many 'empty' columns (with only
zeros). Let us know, if it is a problem.

*This is just some sample data.
DATA LIST LIST /codes (F4).
BEGIN DATA
111
112
113
235
1032
1033
1232
9710
9711
111
112
111
END DATA.

*This is the solution.
COMPUTE two_digt=TRUNC(codes/100).
VECTOR Code (97N1).
DO REPEAT code=Code1 TO Code97 / i=1 TO 97.
COMPUTE code=(two_digt=i).
END REPEAT.
EXE.

Best,
Anton

> -----Original Message-----
> From: SPSSX(r) Discussion [mailto:[hidden email]]
> On Behalf Of Lisa Stickney
> Sent: Tuesday, April 24, 2007 7:45 AM
> To: [hidden email]
> Subject: Recode with part of data in variable name?
>
> Hi SPSSers
>
> I have a dataset with over 1000 cases each containing a
> four digit numeric code ranging from 111 to 9711. I need to
> recode these into a series of dummy variables based on the
> first 2 digits. I would like to combine the name of the code
> with the digit numbers. For example 111 would become Code01
> and 9711 would become Code97. I started doing this as a
> series of IF statements, but it's long & tedious. I know
> there has to be an easier way. Unfortunately I don't know
> how to move the digits into the variable name. Can anyone
> help me? Thanks in advance.
>
> Best,
> Lisa
>
> Lisa T. Stickney
> Ph.D. Candidate
> The Fox School of Business
> and Management
> Temple University
> [hidden email]
>

lts1

Re: Recode with part of data in variable name?

Thanks Anton!

Trust me sparsity is not a problem -- I have few missing codes between the
numbers. I'll give this a run later today and holler back is there's a
problem. Thanks again.

Best,
Lisa

----- Original Message -----
From: "Anton Balabanov" <[hidden email]>
To: "'Lisa Stickney'" <[hidden email]>; <[hidden email]>
Sent: Tuesday, April 24, 2007 3:04 AM
Subject: RE: Recode with part of data in variable name?

> Hi, Lisa,
>
> This syntax is far from general, but goes in your case, is it?
> If your codes are sparse, it produces too many 'empty' columns (with only
> zeros). Let us know, if it is a problem.
>
> *This is just some sample data.
> DATA LIST LIST /codes (F4).
> BEGIN DATA
> 111
> 112
> 113
> 235
> 1032
> 1033
> 1232
> 9710
> 9711
> 111
> 112
> 111
> END DATA.
>
> *This is the solution.
> COMPUTE two_digt=TRUNC(codes/100).
> VECTOR Code (97N1).
> DO REPEAT code=Code1 TO Code97 / i=1 TO 97.
> COMPUTE code=(two_digt=i).
> END REPEAT.
> EXE.
>
>
> Best,
> Anton
>
>> -----Original Message-----
>> From: SPSSX(r) Discussion [mailto:[hidden email]]
>> On Behalf Of Lisa Stickney
>> Sent: Tuesday, April 24, 2007 7:45 AM
>> To: [hidden email]
>> Subject: Recode with part of data in variable name?
>>
>> Hi SPSSers
>>
>> I have a dataset with over 1000 cases each containing a
>> four digit numeric code ranging from 111 to 9711. I need to
>> recode these into a series of dummy variables based on the
>> first 2 digits. I would like to combine the name of the code
>> with the digit numbers. For example 111 would become Code01
>> and 9711 would become Code97. I started doing this as a
>> series of IF statements, but it's long & tedious. I know
>> there has to be an easier way. Unfortunately I don't know
>> how to move the digits into the variable name. Can anyone
>> help me? Thanks in advance.
>>
>> Best,
>> Lisa
>>
>> Lisa T. Stickney
>> Ph.D. Candidate
>> The Fox School of Business
>> and Management
>> Temple University
>> [hidden email]
>>
>
>

lts1

Re: Recode with part of data in variable name?

In reply to this post by Anton Balabanov

Thanks Anton,

This worked like a charm. Now that I look at the results, I have a
handful of codes that have empty columns (all zeros or all zeros & a few
missing values), so the question becomes: without manually going through 97
codes, is there a way to delete the columns that are empty? Thanks in
advance.

Best,
Lisa

----- Original Message -----
From: "Anton Balabanov" <[hidden email]>
To: "'Lisa Stickney'" <[hidden email]>; <[hidden email]>
Sent: Tuesday, April 24, 2007 3:04 AM
Subject: RE: Recode with part of data in variable name?

Anton Balabanov

Re: Recode with part of data in variable name?

Well, than try a bit more complex code:

*Sample data.
DATA LIST LIST /codes (F4).
BEGIN DATA
111
112
113
235
1032
1033
1232
9710
9711
111
112
.
11111
END DATA.

*This is the solution.
COMPUTE two_digt=TRUNC(codes/100).
SORT CASES BY two_digt.
VECTOR Code (97N1).

STRING tempname (A8).
DO IF $casenum=1.
- WRITE OUTFILE="c:/temp/keepvars.sps" /"ADD FILES FILE=* /KEEP=codes ".
END IF.
DO REPEAT code=Code1 TO Code97 / i=1 TO 97.
COMPUTE code=(two_digt=i).
DO IF two_digt=i AND (two_digt~=LAG(two_digt) OR $casenum=1 OR
MISS(LAG(two_digt))).
- COMPUTE tempname=CONCAT(" Code",LTRIM(STRING(i,F4.0))).
- WRITE OUTFILE="c:/temp/keepvars.sps" / tempname .
END IF.
END REPEAT.
EXE.
INCLUDE "c:/temp/keepvars.sps".
exe.

Best,
Anton

> -----Original Message-----
> From: SPSSX(r) Discussion [mailto:[hidden email]]
> On Behalf Of Lisa Stickney
> Sent: Wednesday, April 25, 2007 4:53 PM
> To: [hidden email]
> Subject: Re: Recode with part of data in variable name?
>
> Thanks Anton,
>
> This worked like a charm. Now that I look at the results,
> I have a handful of codes that have empty columns (all zeros
> or all zeros & a few missing values), so the question
> becomes: without manually going through 97 codes, is there a
> way to delete the columns that are empty? Thanks in advance.
>
> Best,
> Lisa
>
>
> ----- Original Message -----
> From: "Anton Balabanov" <[hidden email]>
> To: "'Lisa Stickney'" <[hidden email]>; <[hidden email]>
> Sent: Tuesday, April 24, 2007 3:04 AM
> Subject: RE: Recode with part of data in variable name?
>
>
> > Hi, Lisa,
> >
> > This syntax is far from general, but goes in your case, is it?
> > If your codes are sparse, it produces too many 'empty'
> columns (with
> > only zeros). Let us know, if it is a problem.
> >
> > *This is just some sample data.
> > DATA LIST LIST /codes (F4).
> > BEGIN DATA
> > 111
> > 112
> > 113
> > 235
> > 1032
> > 1033
> > 1232
> > 9710
> > 9711
> > 111
> > 112
> > 111
> > END DATA.
> >
> > *This is the solution.
> > COMPUTE two_digt=TRUNC(codes/100).
> > VECTOR Code (97N1).
> > DO REPEAT code=Code1 TO Code97 / i=1 TO 97.
> > COMPUTE code=(two_digt=i).
> > END REPEAT.
> > EXE.
> >
> >
> > Best,
> > Anton
> >
> >> -----Original Message-----
> >> From: SPSSX(r) Discussion
> [mailto:[hidden email]] On Behalf
> >> Of Lisa Stickney
> >> Sent: Tuesday, April 24, 2007 7:45 AM
> >> To: [hidden email]
> >> Subject: Recode with part of data in variable name?
> >>
> >> Hi SPSSers
> >>
> >> I have a dataset with over 1000 cases each containing a four
> >> digit numeric code ranging from 111 to 9711. I need to
> recode these
> >> into a series of dummy variables based on the first 2 digits. I
> >> would like to combine the name of the code with the digit
> numbers.
> >> For example 111 would become Code01 and 9711 would become
> Code97. I
> >> started doing this as a series of IF statements, but it's long &
> >> tedious. I know there has to be an easier way. Unfortunately I
> >> don't know how to move the digits into the variable name.
> Can anyone
> >> help me? Thanks in advance.
> >>
> >> Best,
> >> Lisa
> >>
> >> Lisa T. Stickney
> >> Ph.D. Candidate
> >> The Fox School of Business
> >> and Management
> >> Temple University
> >> [hidden email]
> >>
> >
> >
>

lts1

Re: Recode with part of data in variable name?

Perfect! The only change I made was to the variable list on the KEEP
subcommand (I included all my other variables).

Thanks again.

Best,
Lisa

----- Original Message -----
From: "Anton Balabanov" <[hidden email]>
To: <[hidden email]>
Sent: Wednesday, April 25, 2007 10:50 AM
Subject: Re: Recode with part of data in variable name?

> Well, than try a bit more complex code:
>
> *Sample data.
> DATA LIST LIST /codes (F4).
> BEGIN DATA
> 111
> 112
> 113
> 235
> 1032
> 1033
> 1232
> 9710
> 9711
> 111
> 112
> .
> 11111
> END DATA.
>
> *This is the solution.
> COMPUTE two_digt=TRUNC(codes/100).
> SORT CASES BY two_digt.
> VECTOR Code (97N1).
>
> STRING tempname (A8).
> DO IF $casenum=1.
> - WRITE OUTFILE="c:/temp/keepvars.sps" /"ADD FILES FILE=* /KEEP=codes ".
> END IF.
> DO REPEAT code=Code1 TO Code97 / i=1 TO 97.
> COMPUTE code=(two_digt=i).
> DO IF two_digt=i AND (two_digt~=LAG(two_digt) OR $casenum=1 OR
> MISS(LAG(two_digt))).
> - COMPUTE tempname=CONCAT(" Code",LTRIM(STRING(i,F4.0))).
> - WRITE OUTFILE="c:/temp/keepvars.sps" / tempname .
> END IF.
> END REPEAT.
> EXE.
> INCLUDE "c:/temp/keepvars.sps".
> exe.
>
> Best,
> Anton
>
>> -----Original Message-----
>> From: SPSSX(r) Discussion [mailto:[hidden email]]
>> On Behalf Of Lisa Stickney
>> Sent: Wednesday, April 25, 2007 4:53 PM
>> To: [hidden email]
>> Subject: Re: Recode with part of data in variable name?
>>
>> Thanks Anton,
>>
>> This worked like a charm. Now that I look at the results,
>> I have a handful of codes that have empty columns (all zeros
>> or all zeros & a few missing values), so the question
>> becomes: without manually going through 97 codes, is there a
>> way to delete the columns that are empty? Thanks in advance.
>>
>> Best,
>> Lisa
>>
>>
>> ----- Original Message -----
>> From: "Anton Balabanov" <[hidden email]>
>> To: "'Lisa Stickney'" <[hidden email]>; <[hidden email]>
>> Sent: Tuesday, April 24, 2007 3:04 AM
>> Subject: RE: Recode with part of data in variable name?
>>
>>
>> > Hi, Lisa,
>> >
>> > This syntax is far from general, but goes in your case, is it?
>> > If your codes are sparse, it produces too many 'empty'
>> columns (with
>> > only zeros). Let us know, if it is a problem.
>> >
>> > *This is just some sample data.
>> > DATA LIST LIST /codes (F4).
>> > BEGIN DATA
>> > 111
>> > 112
>> > 113
>> > 235
>> > 1032
>> > 1033
>> > 1232
>> > 9710
>> > 9711
>> > 111
>> > 112
>> > 111
>> > END DATA.
>> >
>> > *This is the solution.
>> > COMPUTE two_digt=TRUNC(codes/100).
>> > VECTOR Code (97N1).
>> > DO REPEAT code=Code1 TO Code97 / i=1 TO 97.
>> > COMPUTE code=(two_digt=i).
>> > END REPEAT.
>> > EXE.
>> >
>> >
>> > Best,
>> > Anton
>> >
>> >> -----Original Message-----
>> >> From: SPSSX(r) Discussion
>> [mailto:[hidden email]] On Behalf
>> >> Of Lisa Stickney
>> >> Sent: Tuesday, April 24, 2007 7:45 AM
>> >> To: [hidden email]
>> >> Subject: Recode with part of data in variable name?
>> >>
>> >> Hi SPSSers
>> >>
>> >> I have a dataset with over 1000 cases each containing a four
>> >> digit numeric code ranging from 111 to 9711. I need to
>> recode these
>> >> into a series of dummy variables based on the first 2 digits. I
>> >> would like to combine the name of the code with the digit
>> numbers.
>> >> For example 111 would become Code01 and 9711 would become
>> Code97. I
>> >> started doing this as a series of IF statements, but it's long &
>> >> tedious. I know there has to be an easier way. Unfortunately I
>> >> don't know how to move the digits into the variable name.
>> Can anyone
>> >> help me? Thanks in advance.
>> >>
>> >> Best,
>> >> Lisa
>> >>
>> >> Lisa T. Stickney
>> >> Ph.D. Candidate
>> >> The Fox School of Business
>> >> and Management
>> >> Temple University
>> >> [hidden email]
>> >>
>> >
>> >
>>
>