OK. It is not easy for PROC TABULATE. But how about this one ?   proc format; value $ fmt 'Kia'='bold' other='light' ; run; proc tabulate data=sashelp.cars; class make type; classlev make/style={fontweight=$fmt.}; var msrp; table make*{style=<parent>{background=white foreground=black}},type*msrp; run;   ... View more

Great, thank you! ... View more

Thank you for your answers. I guess, @GinaRepole´s solution is the most helpful option for me. With the simple example its pretty easy to reproduce the result. But with my dataset it isnt. The problem is the proc-transpose-step. I tried several ways but i got error messages.  Would be fine, if you could help me again. Here is a snipped of my current dataset (as the result of the work.combo-data-step):   have:   What I want is to transpose it like this:     The work.combo-dataset seems to be sorted. So I tried it with following steps:   proc sort data=kombi; by LFBP_erwtyp3 LF_Typ30 tab HRF_P_Sn_N HRF_P_Sn_Sum; run; proc transpose data=work.kombi out=work.trans prefix=LF_Typ30; var HRF_P_SN_Sum HRF_P_Sn_N; by LFBP_erwtyp3; id LF_Typ30; run;   But the result is for example: "ERROR: The ID value "LF_Typ3001_A_ERZ_1Ki" occurs twice in derselben BY-Gruppe." and "ERROR: All BY groups were bad." Where is the problem of my code? ... View more

data have; input var12 var2 var3 var1 var4; cards; 1 2 3 4 5 ; proc sql noprint; select name into :names separated by ',' from dictionary.columns where upcase(libname)='WORK' and upcase(memname)='HAVE' order by input(compress(name,,'kd'),best.); create table want as select &names. from have; quit; ... View more

Thank you! ... View more

Thank you! ... View more

Perfect, thx for it! ... View more

Great, Thx! ... View more

I am not grasping how the programmer knows which records bet summed together into a group, and which records not to sum, what is the rule? If you want sums ABOVE the rows with the data, you would have to create a data set somehow with everything in the right order including sums above the data to be summed, and then use PROC REPORT on that. ... View more

Great, thank you! ... View more

thank you for your answers. pretty helpful! that works! Ive another question: I guess, proc tabulate doenst know such thing like "select distinct" for a specific column? ... View more

Thank you for your two options. Both are good! ... View more

Each "partei " has only one obs ? you can not use that kind of data to CLUSTER Analysis. For a BY group, you need multiple obs (and better nobs of each group > number of variables) to do CLUSTER. I think you could drop " BY partei;" , that result might be what you are looking for. ... View more