I'm trying to figure out how to specify lengths of new character variables (i.e., that don't already exist in the dataset) whose names are stored in a macro. The way I stored the list of "words" in the macro looks like this: var1$20 var2$30 var3$15 [and so on], in anticipation of using them in a length statement (and format statement). I guess I can't do something like the following (where "len" is the name of the macro that contains the values like above):   %macro x(x); %local i next; %do i=1 %to %sysfunc(countw(&len)); %let next=%scan(&len, &i); data mydata; set mydata; where x=&x; length &next; format &next.; run; %end; %mend x; %x(1) Is there something I can manipulate in this code to make it work, or is there a better way I can specify lengths of new character variables whose names are stored in a macro, where the length would be some numeric value associated with each variable-to-be?    Thanks! ... View more

Hi SAS people,   I'm wondering if there's an easy way (like a function?) to treat an expression as an actual variable/variable name in the creation of a variable in a DATA step. For example, if I have the following data, data have; input id $ first_fla_lang $ first_fla_grd first_flb_lang $ first_flb_grd first_flc_lang $ first_flc_grd second_fla_lang $ second_fla_grd second_flb_lang $ second_flb_grd; cards; 1234567 FRENCH 3.667 SPANISH 4 SPANISH 3 RUSSIAN 2.667 RUSSIAN 3 1234566 SPANISH 4 SPANISH 4 SPANISH 3 PORTUGUESE 3.667 PORTUGUESE 4 1234565 ITALIAN 2 ITALIAN 3 ITALIAN 3 ITALIAN 2.333 GREEK 3 ; run; I would like to be able to have an expression (in this case, using the tranwrd function) act as an actual variable in the DATA step so that the actual value of the newly created variable is the value of the variable being referred to in the expression (which results to the variable name of the variable that I want the new variable to be equal to). So the code would be something like this: data want; set have; x=/*[insert function here?]*/(tranwrd("first_fla_lang","lang","grd")); y=first_fla_grd; run; The result should be that the newly created variables x and y are equal to each other (i.e., equal to the value of the numerical variable first_fla_grd), with the resulting data looking like this: instead of x being a string variable equal to whatever the expression results to, like this (don't want):   Thanks so much!     ... View more

Hi SAS people,   I'm wondering why when I use the SCAN function to extract a substring into a new variable which I then use as the second argument (target) of the TRANSTRN function in a subsequent line of code, the resulting variable created from the TRANSTRN function does not remove/replace the target substring whereas the exact same substring just manually created as a new variable (i.e., without using the SCAN function) does result in the target substring having been removed when it is specified in that second argument of the TRANSTRN function. Unless I specify the exact length of the targeted substring beforehand...   I.e., given this data set   data have; infile datalines truncover; input str $10.; datalines; U-FR/IT-AB U-FRIT-AB U-FREN-AB U-CLST-AB ;     what I would like to see in a resulting data set is the "-AB" at the end of each value for str removed, resulting in "U-FR/IT", "U-FRIT", "U-FREN", and "U-CLST", respectively.    Here are three versions of the code I am using to accomplish this, each with the resulting table. Notice I also create another variable (sub2) with the target substring ("-AB") manually, which I then show is in fact the same exact value by comparing the two in a variable called "diff": data want1 (drop=x); set have; x=compbl(scan(str,3,"-")); sub1=compbl(cats("-",compbl(x))); sub2="-AB"; str1=transtrn(str,sub1,trimn("")); str2=transtrn(str,sub2,trimn("")); if sub1 ne sub2 then diff=1; run;   data want2 (drop=x); set have; length x $4; format x $4.; length sub1 $4; format sub1 $4.; length sub2 $4; format sub2 $4.; x=compbl(scan(str,3,"-")); sub1=compbl(cats("-",compbl(x))); sub2="-AB"; str1=transtrn(str,sub1,trimn("")); str2=transtrn(str,sub2,trimn("")); if sub1 ne sub2 then diff=1; run;   data want3 (drop=x); set have; length x $3; format x $3.; length sub1 $3; format sub1 $3.; length sub2 $3; format sub2 $3.; x=compbl(scan(str,3,"-")); sub1=compbl(cats("-",compbl(x))); sub2="-AB"; str1=transtrn(str,sub1,trimn("")); str2=transtrn(str,sub2,trimn("")); if sub1 ne sub2 then diff=1; run;   So my question is: do I seriously need to specify lengths (and maybe formats) of the variable resulting from the SCAN function if I want to use it properly in a subsequent TRANSTRN function? What if the lengths are variable? Do I need to extract the length of each result and then create new variables for each length? Also, why the heck is the length of $4 in the WANT2 dataset working for the three values of str without a forward slash, but not for the one with the forward slash in row 1?   Any insight into this strange issue (strange to me) would be much appreciated! ... View more