I think the only ways that SAS "guesses" at variables types is through proc import or the import wizard. You could do the conversions in a data step using arrays and a do loop to save a little on the typing. EJ ... View more

Thats interesting I would have thought it was the other way around ... I thought the colon abbreviation is relative new addition to the code base like in version 9.2. I use the colon shortcut all the time in 9.3. EJ ... View more

Not sure its clear what you are asking about. Are you wanting the top 5% of customers from the top 5% of transactions? Or do you want the top 5% of customers based on all transactions. And you didnt not mention how to define the "top". Is it total Money, number of items ordered, number of orders, etc... . What ever the definition is there are many ways to get at what you want. Something like this should help you: proc sql noprint;   create table test as   select make, sum(msrp) as amt format=dollar10.   from sashelp.cars   group by make   order by calculated amt desc   ;   select  sum(amt) , round(sum(amt)*.05,1)  into :tamt, :tpct   from test   ;   Create table out as   select make, amt   from test   where amt > &tpct   ; quit; %put &tamt &tpct; EJ ... View more

Sorry just saw that this was a duplicate post ... EJ ... View more

Can you post the actual code you used? The sort and data step should be enough. EJ ... View more

Thanks I knew I was forgetting something! So I think if you put: call execute('%tab_cat_log(AANTAL_WONINGEN,cd_cat_logement'); instead of just the macro call you might have better results. EJ ... View more

just remember that < 20 would include 15 and so forth by the natural of the way you setup the logic. EJ ... View more

After your message came across it hit me that call symput would not work either because the macro variable would not be available until after the data step concludes. So the best way to go would be to use a format, but if you want to use a macro I would do the following: %macro tab_cat_log(aantal,othervar); length &othervar $ 2;      if &aantal = 1 then &othervar = '02';           else if &aantal = 2 then &othervar = '03';                else if &aantal = 3 then &othervar = '04';                         else if &aantal = 4 then &othervar = '05';                          else if &aantal = 5 then &othervar = '06';          else if &aantal = 6 then &othervar = '07';      else if &aantal = 7 then &othervar = '08';      else if &aantal = 8 then &othervar = '09';      else if &aantal = 9 then &othervar = '10';      else if &aantal < 15 then &othervar = '11';      else if &aantal < 20 then &othervar = '12';      else if &aantal < 25 then &othervar = '13';      else if &aantal < 30 then &othervar = '14';      else if &aantal < 40 then &othervar = '15';      else if &aantal < 50 then &othervar = '16';      else if &aantal < 60 then &othervar = '17';      else if &aantal < 80 then &othervar = '18';      else if &aantal < 100 then &othervar = '19';      else &othervar = '20'; %mend tab_cat_log; %macro start_BG_NIEUW_1_compute(input_table=, output_table=);      data temp1;      set &input_table;      %tab_cat_log(AANTAL_WONINGEN,cd_cat_logement);       run; %mend start_BG_NIEUW_1_compute; Hope this helps! EJ ... View more

The way you are using %let SAS is setting n to the value "AANTAL_WONINGEN" and not to the variable value of AANTAL_WONINGEN in the dataset so that is why it is returning the 'else' value all the time. If you replace the %let line with "call symputx('n',AANTAL_WONINGEN);" I think you get closer to what you want. is right using the format method would probably lead to better results.  Frankly since you already had all the "if ... then" statements typed into the second macro, I dont know why you just didnt use standard datastep "if ...then" to set the values. I dont think you need any macros to accomplish what you have described. Hope this helps! EJ ... View more

You can write an Excel Macro (in VBA or through the recorder) that does the graphic import for you which you could call from within your SAS program. I dont have a specific example of this to provide, but I would create the sas graphics files, create the macro off the created graphic files, then go back and add the macro call to your sas program. Alternatively would be to output the data to the excel file and use Excel to do the graphing. This wont help if the data structure changes a lot though. EJ ... View more

Do you want to remove the row instead? you can conditionally delete observations using if b<=0 then delete. Also you dont need do loop for one action. EJ ... View more