Hi Tom, Thank you for the guidance. As you suggested Proc summary is the best choice. I tried running both proc freq and proc summary. 1. Proc freq - i got the percent frequency count and percentage of total frequency in the output dataset - so if i remove the noprint option and run i am getting the exact number which i am looking out for. 2. proc summary - here i got the type , freq and weight in the output dataset. how can i go about in converting or transpoing the dataset to the exact percentage i am getting in proc freq. Please guide, Thanks and Regards, S.S.Pradeep ... View more

Hi Tom, I need one more guidance on the proc frequency.  I have report to be generated with 250 attributes or variables. in the below code i want the attributes to keep changing till 250 automatically and get the combination. is there any way or do i need to hardcode. proc freq; tables brand*market*aa1*(attrib1)*b5_own /noprint out=freq ; weight weight ; run; What i am currently doing is %macro attrib; proc freq; tables brand*market*aa1*(&attributes)*b5_own /noprint out=freq ; weight weight ; run; %mend attrib; %attrib (attributes= attrib1 attrib2 attrib3.....till attrib250). Can you please guide is there any smarter way to do this? Thanks and Regards, S.S.Pradeep ... View more

Hi Tom, Thank you for the guidance. As you suggested Proc summary is the best choice. I tried running both proc freq and proc summary. 1. Proc freq - i got the percent frequency count and percentage of total frequency in the output dataset - so if i remove the noprint option and run i am getting the exact number which i am looking out for. 2. proc summary - here i got the type , freq and weight in the output dataset. how can i go about in converting or transpoing the dataset to the exact percentage i am getting in proc freq. Please guide, Thanks and Regards, S.S.Pradeep ... View more

Hi Tom, Thank you for the tips. Please find below the code which i am trying.. but unable to get the desire results. I have four variables - 1) b5_own has two values ( 0 ,1)                                       2) AA1 has two values (0 ,1)                                       3) brand ( has 20 brands)                                       4) market ( has 10 markets) I have modified your code %macro twoway(varlist); %local i j k l; %do i=1 %to %sysfunc(countw(&varlist,%str( )))-1; %do j=&i+1 %to %sysfunc(countw(&varlist,%str( ))); %do k=1 %to %sysfunc(countw(&varlist,%str( ))); %do l=1 %to %sysfunc(countw(&varlist,%str( ))); %scan(&varlist,&i)*%scan(&varlist,&j)*%scan(&varlist,&k)*%scan(&varlist,&l) %end; %end; %end; %end; %mend; proc freq data=chks; tables %twoway(b5_own aa1 brand market) ; weight weight; run; Can you please guide where the mistake is because i couldnt match the numbers.. Thanks and Regards, S.S.Pradeep ... View more

Hi Tom, Thank you. As of now its creating all the combinations. i need one more favour. All the 5 variables are binary. for eg. A has value 0 and 1, B has a value 0 and 1 and C.... So what i need an help on is . 0 of A with 1 of B,                                                    1 of A with 0 of B,                                                   1 of A with 1 of B,                                                     0 of A with 0 of B the macro i am looking out for should do for all the A to E variables. Please let me know the thoughts, Thanks and Regards, S.S.Pradeep ... View more

Hi Tom, Thank you .It worked perfectly .I also need one help suppose we have 100 variables then how can we list then inside the macros . its not possible to give all the 100 variables inside the proc freq. In this example we have just 5 variables so we could list them . Can we have generalized code to include all the 100 variables. Can we extend this to a Three way or Four way classifcation? Please let me know your thoughts, Thanks and Regards, S.S.Pradeep Message was edited by: pradeep sridharan Can we ... View more

Hi, Thank you so much for the reply. It worked perfectly. Just would like to know since here we are having only 5 variables. say from A to E . so its easy to assign numbers to them from 1-5. Suppose we have 1000 variables will that be feasible to assign 1000 numbers to all the variables? Please let me know your thoughts Thank you. ... View more

Hi, Thank you for the reply and help. Can we also extend this to more than A B C D E  what i mean is  for 'N' number of variables? Please let me know thanks and regards, S.S.Pradeep ... View more