One correction -

proc sql; select distinct cats('%run_report(',firstname, > > ');') into :macrocall separated by ' ' from meansset; > quit;

should be

proc sql; select distinct cats('%run_report(',firstname, > > ');') into :macrocall separated by ' ' from names;; > quit;

I had renamed it in my SAS copy :)

-Joe

On Fri, Jan 23, 2009 at 2:03 PM, Joe Matise <snoopy369@gmail.com> wrote:

> First off, run your proc means this way: > > %MACRO run_report(firstname_var); > ods rtf file=%outputfolder(filename="&firstname_var..rtf"); > *you need two . here because macro name goes from & to . inclusive if . is > present, the second . is the actual . for the filename; > proc means data=names; > where compress(firstname)= "&firstname_var"; *by is > not appropriate, as it takes a variable, not a variable value, as argument; > var books; > quit; > ods rtf close; > %MEND run_report; > > I'd also change %outputfolder to instead be a macro variable %let > outputfolder=c:\; because there's no reason you need a macro for that. > > Now you just need to do this: > > > proc sql; > select distinct cats('%run_report(',firstname,');') into :macrocall > separated by ' ' from meansset; > *you do not need namelist unless there is something else going on, distinct > will select each name once; > quit; > > > to generate your macro calls, and then > &macrocall > to actually call them (to do the looping). > > Final program: > > > > data names; > input firstname $ books; > cards; > Alice 2 > Alice 4 > Alice 5 > Jim 6 > Jim 3 > Jim 2 > Jim 2 > Jim 4 > Jim 5 > Peter 1 > Peter 6 > Peter 4 > Peter 3 > Peter 7 > ; > run; > > %let outputfolder=C:\; > > > %MACRO run_report(firstname_var); > ods rtf file="&outputfolder.&firstname_var..rtf"; > *you need two . here because macro name goes from & to . inclusive if . is > present, the second . is the actual . for the filename; > proc means data=names; > where compress(firstname)= "&firstname_var"; *by is > not appropriate, as it takes a variable, not a variable value, as argument; > quit; > ods rtf close; > %MEND run_report; > > proc sql; > select distinct cats('%run_report(',firstname,');') into :macrocall > separated by ' ' from meansset; > quit; > > &macrocall; > > Hope this helps, > > Joe > > > On Fri, Jan 23, 2009 at 1:13 PM, webonomic <webonomic@gmail.com> wrote: > >> I've simplified my problem to make it easier. >> >> I have "name_list" dataset which was created from "names" dataset. I >> want to loop through the values of name_list (i.e. Alice, Jim, and >> Andy). For each name, I want to run a PROC MEANS against the "names" >> dataset and spit out an RTF file with the appropriate name. I should >> end up with 3 different files. >> >> This is what I have so far: >> >> data name_list; >> input firstname $; >> cards; >> Alice >> Jim >> Peter >> ; >> run; >> >> data names; >> input firstname $ books; >> cards; >> Alice 2 >> Alice 4 >> Alice 5 >> Jim 6 >> Jim 3 >> Jim 2 >> Jim 2 >> Jim 4 >> Jim 5 >> Peter 1 >> Peter 6 >> Peter 4 >> Peter 3 >> Peter 7 >> ; >> run; >> >> /*outputfolder macro comes form my macro library*/ >> %MACRO outputfolder(filename=); >> "C:\&filename"; >> %MEND outputfolder; >> %global outputfolder; >> >> /*will I get the right filename? or do I need to go &&name_var ?*/ >> %MACRO run_report(firstname_var); >> ods rtf file=%outputfolder(filename="&name_var.rtf"); >> proc means data=names; >> by &firstname_var; >> quit; >> ods rtf close; >> %MEND run_report; >> >> >> /* code to loop through each obs of NAMES_LIST and run RUN_REPORT >> macro */ >> loop through NAMES values; >> for each run %run_report(firstname); >> end loop; >> > >