The class statement has the same role in all regression procs.   A CLASS statement declares a categorical variable for use in your model. Parameters will be fitted for each level of the class variable.   A BY statement fits as many models as there are by-levels. It is a shortcut to submitting the proc for every level of the by-variable.   by sex; model weight = height;   fits two models with separate slopes and interceps.   class sex; model weight = sex height;   fits a single model with two interceps and a single slope.   hth ... View more

proc summary data=have nway; class region; var sales; output out=total_sales sum=total_sales; run; proc sort data=have; by region descending sales; run; data have2; merge have total_sales; by region; if first.region then cumulative_percent=0; percent_of_sales = sales/total_sales; cumulative_percent + percent_of_sales; start_of_grouping = cumulative_percent - percent_of_sales; run; ... View more

Here is an even easier option to clean the data of blank values and then transpose (that doesn't require the macro looping):   DATA WORK.Have (DROP= Type2 Maturity2 Series2); SET WORK.Have; RETAIN Type2 Maturity2 Series2; IF Type NE ' ' THEN Type2=Type; ELSE IF Type = ' ' THEN Type=Type2; IF Maturity NE . THEN Maturity2=Maturity; ELSE IF Maturity = . THEN Maturity=Maturity2; IF Series NE ' ' THEN Series2=Series; ELSE IF Series = ' ' THEN Series=Series2; RUN; PROC SORT DATA=WORK.Have; BY Type Maturity Series ID; RUN; PROC TRANSPOSE DATA=WORK.Have out=WORK.WANT (drop=_name_) prefix=ID_; var ID; by Type Maturity Series; run; ... View more

I generated two tables with some Dummy data:   DATA WORK.Have_Dog; FORMAT Dog_Breed $9. Dog_Gender $6. Dog_Name $12.; INFORMAT Dog_Breed $9. Dog_Gender $6. Dog_Name $12.; INPUT Dog_Breed Dog_Gender Dog_Name; DATALINES; Retriever Male Dog_Retr Labrador Female Dog_Labr Poodle Male Dog_Pood Husky Female Dog_Husk Mutt Male Dog_Mutt ; DATA WORK.Have_Cat; FORMAT Cat_Breed $9. Cat_Gender $6. Cat_Name $12.; INFORMAT Cat_Breed $9. Cat_Gender $6. Cat_Name $12.; INPUT Cat_Breed Cat_Gender Cat_Name; DATALINES; Siamese Male Cat_Siam Persian Female Cat_Pers Savannah Male Cat_Sava Himalayan Female Cat_Hima Mix Male Cat_Mix ;     And this is the logic to perform the task you were requesting: PROC SQL; CREATE TABLE WORK.WANT_AllCombos AS SELECT * FROM WORK.Have_Dog, WORK.HAve_Cat; QUIT;         This is what the results of the dummy data going through this logic to join the data in the two tables: Dog_Breed Dog_Gender Dog_Name Cat_Breed Cat_Gender Cat_Name Retriever Male Dog_Retr Siamese Male Cat_Siam Retriever Male Dog_Retr Persian Female Cat_Pers Retriever Male Dog_Retr Savannah Male Cat_Sava Retriever Male Dog_Retr Himalayan Female Cat_Hima Retriever Male Dog_Retr Mix Male Cat_Mix Labrador Female Dog_Labr Siamese Male Cat_Siam Labrador Female Dog_Labr Persian Female Cat_Pers Labrador Female Dog_Labr Savannah Male Cat_Sava Labrador Female Dog_Labr Himalayan Female Cat_Hima Labrador Female Dog_Labr Mix Male Cat_Mix Poodle Male Dog_Pood Siamese Male Cat_Siam Poodle Male Dog_Pood Persian Female        Cat_Pers Poodle Male Dog_Pood Savannah Male        Cat_Sava Poodle Male Dog_Pood Himalayan Female        Cat_Hima Poodle Male Dog_Pood Mix        Male        Cat_Mix Husky Female Dog_Husk Siamese        Male        Cat_Siam Husky Female Dog_Husk Persian        Female        Cat_Pers Husky Female Dog_Husk Savannah Male        Cat_Sava Husky Female Dog_Husk Himalayan Female        Cat_Hima Husky Female Dog_Husk Mix        Male        Cat_Mix Mutt Male Dog_Mutt Siamese        Male        Cat_Siam Mutt Male Dog_Mutt Persian        Female        Cat_Pers Mutt Male Dog_Mutt Savannah Male        Cat_Sava Mutt Male Dog_Mutt Himalayan Female        Cat_Hima Mutt        Male        Dog_Mutt Mix        Male        Cat_Mix     ... View more

Sounds like you want use the value of macro variable to construct a filename? So if your macro variable has the value 21074 and you want to interpret that as a date and generate a human readable string that you could include as part of the filename you need to use a FORMAT.  I would recommend either YYMMDDD10. (2017-09-12) or YYMMDDN8. (20170912) so that your resulting filenames will sort in chronological order. Perhaps you can apply the format earlier so that the macro variable already looks like the string you want. Or else use the %SYSFUNC() to call the PUTN() function to apply the format to the value.   %let date=21074; %let filename=filename_%sysfunc(putn(&date,yymmddn8.)).xlsx; proc export file="&filename" ....   ... View more