Yes! I kept saying where type = 2 instead of 'char'. Thank you, Hai.Kuo! ... View more

The problem is that I don't know how to modify the code that it tells SAS, hey do this only if I tell you so... I've been trying to use type = "2'...but it doesn't recognize it. data test;   input (x y) ($) z;   cards; 1z400 this1 3 4 45 4 7 . 5 8isnt 0900 6 ; /*add a merge variable for later*/ data test;set test; merge = 1;run; /*use proc contents to get the type of variables, character or numeric*/ proc contents data = test out = t(keep = NAME type);run; proc transpose data = t out = t_t prefix= type_; id name ;var type;run; data t_t;set t_t;merge = 1;run; proc format;     value typ 1 = "Numeric" 2 = "Character"; quit; /*merge with original data set, now we their type*/ data temp(drop = merge);     merge test t_t(drop = _NAME_ _LABEL_);     by merge;     format type_: typ.; /*hard code just so I can see what happens, and because I don't have time to make it general*/     _x = input(x, ?? 32.);     _y = input(y, ?? 32.);     _z = input(z, ?? 32.); /*if the variable of interest is not missing and it is not numeric then do*/ if _x ne . & type_x ne 1 then new_x = "numb";else new_x = x; /*but if it is either missing or numeric, then keep the original value*/ if _y ne . & type_y ne 1 then new_y = "numb";else new_y = y; if _z ne . & type_z ne 1 then new_z = "numb";else new_z = z; run; proc freq data = want;      tables new_:; run; OK, the SAS experts will modify this code into a beautiful piece. I stop here, so I get some work done. Cheers and best of luck! ... View more

Excellent! Now we can all move on! 🙂 ... View more

The problem is the in the bolded line you use lowercase instead of uppercase. You use this proc sql noprint;            select catt("_",name,"=input(",name,",?? 32.);",                "if not missing(_",name,") then do;",                name,"='numb';end;") into :missings            separated by " "                from dictionary.columns                    where libname="work" and memname="test"            ; : Instead of this proc sql noprint;   select catt("_",name,"=input(",name,",?? 32.);",               "if not missing(_",name,") then do;",               name,"='numb';end;") into :missings     separated by " "       from dictionary.columns         where libname="WORK" and  memname="TEST"   ; quit; Please try it and let us know 🙂 ... View more

Arthur's code is great. It didn't run because there is a missing ";" after &missings. data test;   input (x y) ($);   cards; 1z400 this1 4 45 7 . 8isnt 0900 ; proc sql noprint;   select catt("_",name,"=input(",name,",?? 32.);",               "if not missing(_",name,") then do;",               name,"='numb';end;") into :missings     separated by " "       from dictionary.columns         where libname="WORK" and             memname="TEST"   ; quit; data want (drop=_:);   set test;   &missings.; run; proc freq data=want;   tables _all_; run; ... View more

Thank you, Rick! I've been trying to indent the forest plot for months! 🙂 Now it's indented! ... View more