Fw: [Quarantine] Converting international long time-strings to valid SPSS 11 date-formats

[IS-5]

----- Original Message -----
From: "Werner Mueller" <[hidden email]>
To: <[hidden email]>
Sent: Tuesday, December 05, 2006 10:44 AM
Subject: [Quarantine] Converting international long time-strings to valid
SPSS 11 date-formats

> Dear list,
>
> I am having trouble converting a string variable (A45) to a SPSS date
> format.
> It is longer as usual, because it has a leading weekday-name
> AND I have different .sav files from different countries,
> and each weekday-name and month-name is localized!
>
> So I have  dutch  strings
> woensdag 29 november 2006 20:48:59
> woensdag 29 november 2006 20:54:54
> woensdag 29 november 2006 20:56:02
> woensdag 29 november 2006 20:58:48
> zondag 3 december 2006 11:51:14
> zondag 3 december 2006 12:19:40
> zondag 3 december 2006 12:31:26
>
> and also spanish strings
> lunes, 27 de noviembre de 2006 14:26:20
> Martes, 28 de Noviembre de 2006 02:30:58 a.m.
> lunes, 27 de noviembre de 2006 14:38:55
> martes, 28 de noviembre de 2006 14:41:43
> martes, 28 de noviembre de 2006 15:13:18
> Martes, 28 de Noviembre de 2006 03:22:16 a.m.
> lunes, 27 de noviembre de 2006 15:33:39
> martes, 28 de noviembre de 2006 15:31:47
> lunes, 27 de noviembre de 2006 15:49:17
>
> and also french, italian,  ....
> and, some have the 24-hour format, some work with a.m./p.m. (as you can
> see above).
>
> 1. I would like to convert them all to the same time format and
>   because I have 3 of those 45-Character String-Variables for each
> resondent  (time1, time2, time3),
> 2. I would like to calculate differences between time1, time2, and time3.
>   Usually all three times belong to one day, and it would be OK,
>   if I loose the dayname and just calculate the differences in
> hours/minutes.
>
> I searched in the last  20000 emails from this group and also
> Mr.Levesque's homepage,
> but I did not find any time-conversion-examples that dealed with those
> long time-stamps
> and the differrent country-languages + partial am/pm problems.
>
> I confess, I come undone with this problem!
>
> Any help would be much appreciated.
>
> Werner Mueller
> (University of Duesseldorf, Germany)

[hillel vardi]

Shalom


Here is another way of doing what you need . it will run on spss 11.5
(needs the casetovars)  . with a small adaptation it can run even on
spss 6 .

The advantage of this approach  is using aggregation witch ensure  that
you will have all the data needed for the translation (recoding) . also
it is relatively small.

The program assume that all data have the same order (weekday day month
year time am_pm) .


DATA LIST FREE/ str(a60).
BEGIN DATA
 zz1  woensdag 29 november 2006 20:48:59
 zz2 woensdag 29 november 2006 20:54:54
 zz3 woensdag 29 november 2006 20:56:02
 zz4 woensdag 29 november 2006 20:58:48
 zz5 zondag 3 december 2006 11:51:14
 zz6 zondag 3 december 2006 12:19:40
 zz7 zondag 3 december 2006 12:31:26
 zz8 lunes, 27 de noviembre de 2006 14:26:20
 zz9 Martes, 28 de Noviembre de 2006 02:30:58 a.m.
 zz10 lunes, 27 de noviembre de 2006 14:38:55
 zz11 martes, 28 de noviembre de 2006 14:41:43
 zz12 martes, 28 de noviembre de 2006 15:13:18
 zz13 Martes, 28 de Noviembre de 2006 03:22:16 a.m.
 zz14 lunes, 27 de noviembre de 2006 15:33:39
 zz15 martes, 28 de noviembre de 2006 15:31:47
 zz16 lunes, 27 de noviembre de 2006 15:49:17
end data.
compute   seq=$casenum.
sort cases   by str .
SAVE OUTFILE='date_data.sav'  /COMPRESSED.

aggregate  outfile= * / break=str / n=n(str) .
select if    (substr(str,1,1) > '9') and (substr(str,1,2) < 'zz' ) .

SAVE OUTFILE='date_hlp.sav'  /COMPRESSED.


*********************************************************** .

*   After you do the aggregate you should  add the new(translate)
values  to the help table(data_hlp.sav)  to get a table like this  .

*  (datetype are  2= weekday 3=month  0=discard 1=am_pm  )  .

*  I am adding the table as data here because  there is no way to attach
the created file .

*********************************************************** .


data list list /str(a20) n datetype dataval .
BEGIN DATA
a.m.                  2      1
de                   18      0
december              3      2       12
lunes                 4      3        2
martes                3      3        3
Martes                2      3        3
november              4      2       11
noviembre             7      2       11
Noviembre             2      2       11
woensdag              4      3        4
zondag                3      3        1
end data.

*match files    file='date_data.sav'
               / table='date_hlp.sav'  .
match files    file='date_data.sav'
               / table= *
              / by str/ drop=n .

select if     sysmis(datetype) or datetype gt 0 .
sort cases    by seq.
numeric        caseseq(f1)  casenum(f8).
string        newstr(a20) .
leave          caseseq  casenum newstr.
if            substr(str,1,2) eq 'zz' caseseq=0.
if            substr(str,1,2) eq 'zz' casenum=sum(casenum,1).
compute       caseseq=sum(caseseq,1) .
compute       caseseq=caseseq .
compute       casenum=casenum .
compute       newstr=str.
if            datetype eq 2 or datetype eq 3  newstr=strng(dataval,f4) .
execute .
execute .
SORT CASES BY casenum caseseq .
formats    caseseq(f1).
CASESTOVARS
 /ID = casenum
 /INDEX = caseseq
 /drop=str seq datetype dataval
  /GROUPBY = VARIABLE .
rename vars    (newstr.1 newstr.2 newstr.3 newstr.4 newstr.5 newstr.6
newstr.7
             = idcode weekday day month year time am_pm ).

compute
spssdate=date.dmy(number(ltrim(tday),f2),number(ltrim(tmonth),f2),number(ltrim(tyear),f4))
.
compute     spsstime=TIME.HMS(number(substr(ltrim(ttime),1,2),f2)
                             ,number(substr(ltrim(ttime),4,2),f2)
                             ,number(substr(ltrim(ttime),7,2),f2)  ).

if          index(lower(am_pm),'pm')  spsstime=spsstime + (12*60*60) .

formats     spssdate(edate12) spsstime(time8).


execute .



Hillel Vardi




IS wrote:

> ----- Original Message -----
> From: "Werner Mueller" <[hidden email]>
> To: <[hidden email]>
> Sent: Tuesday, December 05, 2006 10:44 AM
> Subject: [Quarantine] Converting international long time-strings to valid
> SPSS 11 date-formats
>
>
>> Dear list,
>>
>> I am having trouble converting a string variable (A45) to a SPSS date
>> format.
>> It is longer as usual, because it has a leading weekday-name
>> AND I have different .sav files from different countries,
>> and each weekday-name and month-name is localized!
>>
>> So I have  dutch  strings
>> woensdag 29 november 2006 20:48:59
>> woensdag 29 november 2006 20:54:54
>> woensdag 29 november 2006 20:56:02
>> woensdag 29 november 2006 20:58:48
>> zondag 3 december 2006 11:51:14
>> zondag 3 december 2006 12:19:40
>> zondag 3 december 2006 12:31:26
>>
>> and also spanish strings
>> lunes, 27 de noviembre de 2006 14:26:20
>> Martes, 28 de Noviembre de 2006 02:30:58 a.m.
>> lunes, 27 de noviembre de 2006 14:38:55
>> martes, 28 de noviembre de 2006 14:41:43
>> martes, 28 de noviembre de 2006 15:13:18
>> Martes, 28 de Noviembre de 2006 03:22:16 a.m.
>> lunes, 27 de noviembre de 2006 15:33:39
>> martes, 28 de noviembre de 2006 15:31:47
>> lunes, 27 de noviembre de 2006 15:49:17
>>
>> and also french, italian,  ....
>> and, some have the 24-hour format, some work with a.m./p.m. (as you can
>> see above).
>>
>> 1. I would like to convert them all to the same time format and
>>   because I have 3 of those 45-Character String-Variables for each
>> resondent  (time1, time2, time3),
>> 2. I would like to calculate differences between time1, time2, and
>> time3.
>>   Usually all three times belong to one day, and it would be OK,
>>   if I loose the dayname and just calculate the differences in
>> hours/minutes.
>>
>> I searched in the last  20000 emails from this group and also
>> Mr.Levesque's homepage,
>> but I did not find any time-conversion-examples that dealed with those
>> long time-stamps
>> and the differrent country-languages + partial am/pm problems.
>>
>> I confess, I come undone with this problem!
>>
>> Any help would be much appreciated.
>>
>> Werner Mueller
>> (University of Duesseldorf, Germany)
>