counting observations depending on two variables

[lone0708]

Hi,

I have a dataset that look like the following:

pt_id   var.  var. var. date_examination       date_group

1.                                   01.01.90                      1

1                                    02.01.90                      1

1                                    03.01.90                      2

2                                    01.01.90                      1

3                                    03.01.90                      4

3                                    10.01.90                      3

I want to count how many patients have examination dates in the specific date_groups but only appear 1 time in the output. Some patients have several examinations in the same date_group.

Any ideas?

Thanks 🙂

[ballardw]

You might end up having to provide a bit longer example. If variables do not have any affect on the output please do not bother to include them in example data.
So do your 3 "var." variables have any role? or the date_examination? If so describe what that might be.

I can interpret your description more than one way so you may end up having to provide which desired output.

One way is to count the combinations of pt_id and date_group and then count the pt_id in that result. The second result gives how many times pt_id appeared with different date group(s) but not which ones.

Proc freq data=have noprint;
   tables pt_id*date_group/list out=count1 nocum;
run;

proc freq data=count1;
  tables pt_id;
run;

It is best to provide example data in the form of a data step so we can recreate your data. Post the code in a code box opened on the forum with the </> icon as I have to  prevent the message window from reformatting the text so that the code will not run.

[lone0708]

Hi Ballardw,
thanks for your reply, I am sorry for the poor question.
The 3 "var" has no role. The date_examination was to illustrate, that dates can vary, but the date_group does not necessarily change because of that.
I can't provide example data, as I am not allowed to.
Your code works, but is it possible to see/add which date_groups the pt_id appears in?

[ballardw]

---

@lone0708 wrote:
Hi Ballardw,
thanks for your reply, I am sorry for the poor question.
The 3 "var" has no role. The date_examination was to illustrate, that dates can vary, but the date_group does not necessarily change because of that.
I can't provide example data, as I am not allowed to.
Your code works, but is it possible to see/add which date_groups the pt_id appears in?

---

The count1 data set will have one record per Id date_group. You could print that.

Or either of these for different display than proc print:

proc tabulate data=count1;
   class pt_id date_group;
   table pt_id='',
         date_group*n=''
         /misstext=' ' box=pt_id
   ;
run;

proc report data=count1;
   columns pt_id date_group ;
   define pt_id /group;
   define date_group/ across;
run;

These two tables will have a 1 indicating that the pt_id appeared with that date_group

Or you could use the original data set with either of these to create a matrix with the cells having the count of times the date_group appeared with the Pt_id.