Your post: @Sonia2 wrote: I got the same question in exam but I want to clear some concepts, In same question can I use if in1 and in2 then output both; if in1 and not in2 then output no_grad_year; If not in1 and in2 then output no_income ; And my second question is when we give answer to question in exam if we look at else if in1 and not in2 then output no_grad_year; in output we will get some missing values if I want answer to question and answer is missing value then I’ll write for or type missing value or what ?? Please explain me. Please notice that IN1 and IN2 - each contains 1 for EXIST and 0 for absent.    As long as you didn't arrive to RETURN statement or to the end of the program, the data is still in the memory and you can write it to output, either the same dataset (making duplicates) or to other dataset.   Missing value in a variable is the result of nonexistance record on the dataset that contains this variable.       ... View more

new_var = compress(var); ... View more

As you assigning new values to variables in ALL observations, you don't need to assign missing values to them. Anyhow yoy can control order of variables to show when you open the dataset by using FORMAT statement with or without a format, as in:   DATA NEW(DROP=make); FORMAT <1st_var> <2nd_var> <3rd_var> ... ; SET SASHELP.cars; Type='All'; Horsepower=1000; RUN; ... View more

Replace line: _vars(i) = catt(substr(vname(_vars), 1, 1), _vars(i)); into: _vars(i) = catt(substr(vname(_vars(i)), 1, 1), _vars(i)); ... View more

1) What do you mean by "going mistake" - is there a syntax error or mis calculating? 2) In order to help[ you we need the full code run, better shown in the log within the messegaes. ... View more

I would split your issue into few steps  to deal with: 1) Choose the desired type (NUM/CHAR) of the variable named in the message. 2) Check each data set does it fit the desired type and make a list of data sets that need to change the variable type. Use any proposed method.  3) use next code to cahnge type of a variable:   data correct; set ds_in(rename=(var=_var)); /* adapt variable name */ var = input(_var, best8.2); /* char to num, adapt the format */ /* or */ var = put(_var, 8.2); /* num to char, adapt the format */ drop _var; run; 4) Check the code with a just one data set from the list. You can use      options obs=10; for the test.   5) Use next macro to loop over the data sets that need repair type:   %let list = dsn1 dsn2 dsn3 ...; /* List of data set names to repair */ %macro repair; %do i=1 to %sysfunc(countw(&list)); %let dsn = %scan(&list,&i); data &dsn; set &dsn((rename = (var=_var)); /* enter here the correct conversion statement */ run; %end; %mend repair; %repair; 6) Run your merge code with the full list.       Have a test to ensure you repaired all data sets and all variables using:         set &dsn (obs=10 rename=(var=_var));     in the above macro.  7)  Use options obs=max;  to do it over the whole data set and drop      obs=10 from the macro statement.   ... View more