Yes, your missing will be if disp3_1=. then disp3_1=0 or you could use if nmiss(disp3_1) then disp3_1=0; Two examples are below, but the second is easier -- with the sum (of ...) you don't have to worry about the missing values.

*---- example 1 ---------------------------------------------------*; data a; disp3_1=.; disp3_2=1; disp4_1=2; disp4_2=5; output; disp3_1=1; disp3_2=1; disp4_1=3; disp4_2=6; output; disp3_1=.; disp3_2=.; disp4_1=4; disp4_2=7; output; disp3_1=.; disp3_2=1; disp4_1=5; disp4_2=.; output; disp3_1=5; disp3_2=4; disp4_1=3; disp4_2=1; output; run;

data arth(drop=i j); set a; x=0;

array disp3_{*} disp3_1 disp3_2; array disp4_{*} disp4_1 disp4_2;

do i=1 to dim(disp3_); if nmiss(disp3_{i}) then disp3_{i}=0; x=disp3_{i} + x; end;

do j=1 to dim(disp4_); if nmiss(disp4_{j}) then disp4_{j}=0; x=disp4_{j} + x; end;

run;

proc print data=arth; run;

*--- example 2 --------------------------*;

data sumit; set a; array values[*] disp3_1 disp3_2 disp4_1 disp4_2; tally=sum(of values[*]); run;

proc print data=sumit; run;

On 12/5/06, jshansen <selchauhansen@gmail.com> wrote: > > On 5 Dec., 12:15, "Priya" <priyawor...@yahoo.com> wrote: > > Hi Every1, > > > > I have a small difficulty in writing the assignment statements in an > > efficient manner. > > > > For eg, in a dataset I have variables like DISP3_1 DISP3_2 DISP4_1 > > DISP4_3, etc > > > > If have to do convert all missing values for DISP eq to zero, how will > > I do it without using do loop and without having to manually write all > > assignment statements like > > > > if DISP3_1 = "." then DISP3_1 = 0; > > if DISP3_2 = "." then DISP3_2 = 0; > > and so on ... > > > > Writing a do loop may not be a good of doing this if I do not want to > > track all the display variables created in the dataset which may be > > huge. > > > > I tried doing something like: > > if DISP: = "." then DISP: = 0; > > > > but it did not work. > > > > Kindly suggest a way to handle it in a smart manner.(This may hold true > > for any arithmetic operation.) > > > > Also, suggest if we can convert all missing values for all the > > variables in a dataset to zero in one go. > > > > Thanks, > > Priya > > Priya ! > > A missing value is NOT enclosed in quotes !! > Nor are any numeric values. > > Mind You than you should NOT do this, unless you have a very very good > reason to do it. > > data a; > set sashelp.class; > array numbers _numeric_; > do over numbers; > numbers=sum(numbers,0); > end; > run; > > Jan Selchau-Hansen >