creating string variable from a numberic variable

[Arthur Kramer]

I have a data file with numeric date fields that I would like to change to string variables so I can create substrings. For example, a date like 20120825 I would like to break out to three variables:  2012, another variable 08, and another 25.

 

Thanks for any help you can provide.

 

 

Arthur Kramer, PhD

Director of Institutional Research

New Jersey City University

 

 

 

[Bruce Weaver]

Hi Art.  See if this example helps.  It assumes the date strings are all 8 characters in length, as in the example you gave.  If that is not so, the syntax will have to be made more general.

* Read in some sample data.
new file.
dataset close all.
data list list / StringVar (A8).
begin data
20120825
20120926
20121027
end data.

* The following uses some code borrowed from
* the Date & Time Wizard.

* Pull out YYYY-MM-DD components.
* Here, I assume you actually want to store them as separate variables.
COMPUTE Year =  number(substr(ltrim(StringVar),1,4),f4.0).
COMPUTE Month = number(substr(ltrim(StringVar),5,2),f2.0).
COMPUTE Day = number(substr(ltrim(StringVar),7,2),f2.0).
* I assume you really want a true date variable at the end of the day.
COMPUTE DateVar = date.dmy(Day,Month,Year).
FORMATS DateVar (DATE11).
VARIABLE WIDTH  DateVar(11).
LIST.

* If you don't need the YYYY-MMM-DD components, you
* could use this syntax pasted from the Date & Time Wizard
* but with things lined up in a way that makes it
* easier to follow (when viewed in fixed font, at least).

* Return file to original state, then re-compute with scratch variables.
DELETE VARIABLES Year to DateVar.

* Date and Time Wizard: DateVar.
COMPUTE DateVar=
 date.dmy(number(substr(ltrim(StringVar),7,2),f2.0),
          number(substr(ltrim(StringVar),5,2),f2.0),
          number(substr(ltrim(StringVar),1,4),f4.0)).
VARIABLE LABELS DateVar ''.
VARIABLE LEVEL  DateVar (SCALE).
FORMATS DateVar (DATE11).
VARIABLE WIDTH  DateVar(11).
LIST.

Arthur Kramer wrote

I have a data file with numeric date fields that I would like to change to string variables so I can create substrings. For example, a date like 20120825 I would like to break out to three variables:  2012, another variable 08, and another 25.

Thanks for any help you can provide.


Arthur Kramer, PhD
Director of Institutional Research
New Jersey City University

[Richard Ristow]

At 05:33 PM 10/18/2013, Arthur Kramer wrote:

>I have a data file with numeric date fields that I would like to
>change to string variables so I can create substrings. For example,
>a date like 20120825 I would like to break out to three
>variables:  2012, another variable 08, and another 25.

If your dates are numerical in that form, you have several choices.

a.) Strings, as you requested:

STRING  Str_DATE (A8).
COMPUTE Str_Date=STRING(Date,F8).

In recent versions of SPSS, you can also use the ALTER TYPE command.

b.) You can break out year, month, and day from the numeric form:


NUMERIC Year (F4),
         Month(F2),
         Day  (F2).

COMPUTE Day   = MOD (Date,100).
COMPUTE Month = MOD((Date-Day)/100,100).
COMPUTE Year  =     (Date-100*Month-Day)/1E4.

c.) Recommended: convert your dates to SPSS date variables, then use
XDATE functions as needed. Proceed as above, preferably using scratch
variables for month, day, and year (i.e., #Year, #Month, #Day), and

COMPUTE #Day   = MOD (Date,100).
COMPUTE #Month = MOD((Date-#Day)/100,100).
COMPUTE #Year  =     (Date-100*#Month-#Day)/1E4.

NUMERIC SPSSDate (DATE11).
COMPUTE SPSSDate =DATE.DMY(#Day,#Month,#Year).

===========================================================
APPENDIX 1: Results of all of the above computations:
|-----------------------------|---------------------------|
|Output Created               |18-OCT-2013 18:14:09       |
|-----------------------------|---------------------------|
     Date Str_DATE Year Month Day    SPSSDate

17760704 17760704 1776    7    4 04-JUL-1776
19440606 19440606 1944    6    6 06-JUN-1944
20131018 20131018 2013   10   18 18-OCT-2013

Number of cases read:  3    Number of cases listed:  3
=================================
APPENDIX 2: Test data, and all code
=================================
NEW FILE.
DATA LIST FREE /Date (F8).
BEGIN DATA
17760704 19440606 20131018
END DATA.

* a.) Strings, as you requested:      ... .

STRING  Str_DATE (A8).
COMPUTE Str_Date=STRING(Date,F8).

*b.) You can break out year, month, and day from the numeric form:

NUMERIC Year (F4),
         Month(F2),
         Day  (F2).

COMPUTE Day   = MOD (Date,100).
COMPUTE Month = MOD((Date-Day)/100,100).
COMPUTE Year  =     (Date-100*Month-Day)/1E4.

*c.) Recommended: convert your dates to SPSS date variables ... .

COMPUTE #Day   = MOD (Date,100).
COMPUTE #Month = MOD((Date-#Day)/100,100).
COMPUTE #Year  =     (Date-100*#Month-#Day)/1E4.

NUMERIC SPSSDate (DATE11).
COMPUTE SPSSDate =DATE.DMY(#Day,#Month,#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

[Albert-Jan Roskam]

> On Saturday, October 19, 2013 12:02 AM, Bruce Weaver <[hidden email]> wrote:

> > Hi Art.  See if this example helps.  It assumes the date strings are all 8
> characters in length, as in the example you gave.  If that is not so, the
> syntax will have to be made more general.
>
> * Read in some sample data.
> new file.
> dataset close all.
> data list list / StringVar (A8).
> begin data
> 20120825
> 20120926
> 20121027
> end data.
>
> * The following uses some code borrowed from
> * the Date & Time Wizard.
>
> * Pull out YYYY-MM-DD components.
> * Here, I assume you actually want to store them as separate variables.
> COMPUTE Year =  number(substr(ltrim(StringVar),1,4),f4.0).
> COMPUTE Month = number(substr(ltrim(StringVar),5,2),f2.0).
> COMPUTE Day = number(substr(ltrim(StringVar),7,2),f2.0).
> * I assume you really want a true date variable at the end of the day.
> COMPUTE DateVar = date.dmy(Day,Month,Year).
> FORMATS DateVar (DATE11).
> VARIABLE WIDTH  DateVar(11).
> LIST.
>
> * If you don't need the YYYY-MMM-DD components, you
> * could use this syntax pasted from the Date & Time Wizard
> * but with things lined up in a way that makes it
> * easier to follow (when viewed in fixed font, at least).

Agree, so using scratch variables instead might be a good compromise:

COMPUTE #Year =  number(substr(ltrim(StringVar),1,4),f4.0).
COMPUTE #Month = number(substr(ltrim(StringVar),5,2),f2.0).
COMPUTE #Day = number(substr(ltrim(StringVar),7,2),f2.0).
COMPUTE DateVar = date.dmy(#Day, #Month, #Year).
FORMATS DateVar (SDATE10).  /* the lesser known sortable date, ISO, but still SPSS-ian.

=====================
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]

At 07:40 AM 10/19/2013, Albert-Jan Roskam wrote:

>[Calculation] using scratch variables instead might be a good compromise:
>
>COMPUTE #Year =  number(substr(ltrim(StringVar),1,4),f4.0).
>COMPUTE #Month = number(substr(ltrim(StringVar),5,2),f2.0).
>COMPUTE #Day = number(substr(ltrim(StringVar),7,2),f2.0).
>COMPUTE DateVar = date.dmy(#Day, #Month, #Year).
>
>FORMATS DateVar (SDATE10).  /* the lesser known sortable date, ISO,
>but still SPSS-ian.

That could give the impression that SPSS-form dates with other
formats are *not* sortable. Changing the format from, say, DATE11. to
SDATE10., changes how the values are *displayed*, but not how they
are represented internally. All SPSS-form dates are sortable.

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

[Arthur Kramer]

Thank you, everyone, for the assistance. For now the most expedient is the simple syntax Richard provided.  I will try the more sophisticated and elegant  syntax at a later date--right now, expedience is the "thing", unfortunately.

Arthur

Arthur Kramer, PhD
Director of Institutional Research
New Jersey City University

-----Original Message-----
From: SPSSX(r) Discussion [mailto:[hidden email]] On Behalf Of Albert-Jan Roskam
Sent: Saturday, October 19, 2013 7:40 AM
To: [hidden email]
Subject: Re: creating string variable from a numberic variable

> On Saturday, October 19, 2013 12:02 AM, Bruce Weaver <[hidden email]> wrote:

> > Hi Art.  See if this example helps.  It assumes the date strings are
> > all 8
> characters in length, as in the example you gave.  If that is not so,
> the syntax will have to be made more general.
>
> * Read in some sample data.
> new file.
> dataset close all.
> data list list / StringVar (A8).
> begin data
> 20120825
> 20120926
> 20121027
> end data.
>
> * The following uses some code borrowed from
> * the Date & Time Wizard.
>
> * Pull out YYYY-MM-DD components.
> * Here, I assume you actually want to store them as separate variables.
> COMPUTE Year =  number(substr(ltrim(StringVar),1,4),f4.0).
> COMPUTE Month = number(substr(ltrim(StringVar),5,2),f2.0).
> COMPUTE Day = number(substr(ltrim(StringVar),7,2),f2.0).
> * I assume you really want a true date variable at the end of the day.
> COMPUTE DateVar = date.dmy(Day,Month,Year).
> FORMATS DateVar (DATE11).
> VARIABLE WIDTH  DateVar(11).
> LIST.
>
> * If you don't need the YYYY-MMM-DD components, you
> * could use this syntax pasted from the Date & Time Wizard
> * but with things lined up in a way that makes it
> * easier to follow (when viewed in fixed font, at least).

Agree, so using scratch variables instead might be a good compromise:

COMPUTE #Year =  number(substr(ltrim(StringVar),1,4),f4.0).
COMPUTE #Month = number(substr(ltrim(StringVar),5,2),f2.0).
COMPUTE #Day = number(substr(ltrim(StringVar),7,2),f2.0).
COMPUTE DateVar = date.dmy(#Day, #Month, #Year).
FORMATS DateVar (SDATE10).  /* the lesser known sortable date, ISO, but still SPSS-ian.

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