topic Re: Split a variable and assign the contribution of IDs to that variable values in New SAS User

Split a variable and assign the contribution of IDs to that variable values

NewUsrStat — Thu, 21 Nov 2024 13:00:59 GMT

Hi guys,

suppose to have the following:

data DB;
  input ID :$20. Admission :date09. Discharge :date09. Index Age_class; 
  format Admission date9. Discharge date9.;
cards;
0001 13JAN2015 20JAN2015 1  6
0001 21FEB2015 31DEC2015 0  6
0001 01MAR2018 30SEP2018 0  7
0001 01JAN2019 31DEC2019 0  7
0002 01JAN2015 31DEC2015 1  3
0002 01JAN2019 31OCT2019 0  5
0003 08FEB2014 10MAR2014 1  3
0003 16JUN2015 13JUL2015 0  3
0004 04MAY2016 10MAY2016 1  8
0004 13SEP2017 15NOV2017 0  8
0004 09DEC2018 31DEC2018 0  8
;

Age_class takes values from 1 to 8.

Is there a way to get the following?

data DB;
  input ID :$20. Admission :date09. Discharge :date09. Index Age_class Age_class1 Age_class2 Age_class3 .... Age_class7 Age_class8; 
  format Admission date9. Discharge date9.;
cards;
0001 13JAN2015 20JAN2015 1  6  1  1  1 ... 1  1
0001 21FEB2015 31DEC2015 0  6  0  0  0 ... 0  0
0001 01MAR2018 30SEP2018 0  7  0  0  0 ... 0  0
0001 01JAN2019 31DEC2019 0  7  0  0  0 ... 0  0
0002 01JAN2015 31DEC2015 1  3  1  1  0 ... 1  1
0002 01JAN2019 31OCT2019 0  5  0  0  0 ... 0  0
0003 08FEB2014 10MAR2014 1  3  1  1  0 ... 1  1
0003 16JUN2015 13JUL2015 0  3  0  0  0 ... 0  0
0004 04MAY2016 10MAY2016 1  8  1  1  1 ... 1  0
0004 13SEP2017 15NOV2017 0  8  0  0  0 ... 0  0
0004 09DEC2018 31DEC2018 0  8  0  0  0 ... 0  0
;

In other words I need to fill Age_class* variables (Age_class1, Age_class2, ..., Age_class8) with 1 if at Index = 1 the patient IS NOT in that age_class and 0 otherwise.

Note that Index = 1 is always the first date (Admission-Discharge) for each patient and there is always one Index = 1 for each patient.

For example: pts 0001 at Index = 1 has Age_class = 6. Age_class1, Age_class2, Age_class3, Age_class4, Age_class5, Age_class7, Age_class8 will be filled with 1 because the patient is not contributing to that age_class while it will be 0 at Age_class6 because it is contributing.

Note also, that there are other dates that will be put to 0 because of no interest.

I know, I can do "if then else" but it will take a lot of coding to perform all comparisons.

Thank you very much for your help.

Re: Split a variable and assign the contribution of IDs to that variable values

PeterClemmensen — Thu, 21 Nov 2024 13:15:54 GMT

Try this

data want(drop = i);
   set DB;
   array a {8} Age_class1 - Age_class8;
   do i = 1 to dim(a);
      a[i] = Index = 1;
      if Index = 1 then a[Age_class] = 0;
   end;
run;

Re: Split a variable and assign the contribution of IDs to that variable values

PaigeMiller — Thu, 21 Nov 2024 13:23:38 GMT

Maybe I'm not understanding this properly, because the admission dates and discharge dates are not used anywhere. Also maybe I'm not understanding the general problem of creating these zero and 1 variables, why do you need these if you have the value of INDEX? Seems redundant to me, what is the next step after you have these 0 and 1 variables, what will you do with them, what analysis or plot or table are you going to create?

Re: Split a variable and assign the contribution of IDs to that variable values

NewUsrStat — Thu, 21 Nov 2024 13:59:44 GMT

Thank you very much for your comment. Yes it is true that there's the variable Age_class but I need to count every time every ID is not contributing to an age class from 1 to 8 except the one to which it belongs. Yes, dates will not be used but are there and I left them to show that Index = 1 at first date.

Re: Split a variable and assign the contribution of IDs to that variable values

PaigeMiller — Thu, 21 Nov 2024 14:04:59 GMT

@NewUsrStat wrote:
Thank you very much for your comment. Yes it is true that there's the variable Age_class but I need to count every time every ID is not contributing to an age class from 1 to 8 except the one to which it belongs. Yes, dates will not be used but are there and I left them to show that Index = 1 at first date.

Far too vague of an explanation about what you will do next with this data.