PROC SQL to SAS DATA STEP

[woo]

hello friends - if you can tell me how can i write below proc sql query in data step? thank you

PROC SQL;
CREATE TABLE WORK.Query2_for_TEST3 AS SELECT DISTINCT TEST3.date,
TEST3.Item_number,
(COUNT(TEST3.found)) AS COUNT_OF_found
FROM WORK.TEST3 AS TEST3
GROUP BY TEST3.Item_number
ORDER BY TEST3.Item_number;
QUIT;

[ballardw]

It would better to show start data and desired end.

DATA step probably wouldn't be the best as it has not equivalent of GROUP BY or DISTINCT.

I would try if found is numeric:

proc summary data = work.test3 nway;

   class Item_number date;

   var found;

   output out=qork.query2_for_test3 (drop= _type_ _freq_) n= count_of_found;

run;

Or did you need group by Item_number, date for your original query?

[woo]

so have this: date (in date format) (item_number and found both are numeric)

date          item_number   found
01oct14    123456            1

02oct14    123456            2

01oct14    789101            1

02oct14    789101            2

03oct14    789101            3

04oct14    789101            4

01oct14    111111            1

01oct14    222222            1

02oct14    222222            2

03oct14    222222            3

and want this

date         item_number    count_of_found
01oct14    123456              2

02oct14    123456              2

01oct14    789101              4

02oct14    789101              4

03oct14    789101              4

04oct14    789101              4

01oct14    111111              1

01oct14    222222              3

02oct14    222222              3

03oct14    222222              3

[Astounding]

A DATA step might look like this:

data want;

   count_of_found=0;

   do until (last.item_number);

      set have;

      by item_number notsorted;

      count_of_found + 1;

   end;

   do until (last.item_number);

      set have;

      by item_number notsorted;

      output;

   end;

run;

Optionally, you could drop FOUND from the final data set.

Good luck.

[LinusH]

Why?