Let's work backwards here.   First thing is that the VALUES clause wants actual VALUES and not EXPRESSIONS. So if C,D and E are NUMERIC variables then do this  The number of decimal places in the value inserted makes no difference.  .80 and .8 are the exact same number. insert into eff_all2(Cohort_new, N, c, d,e) values ("Total","&ne_allt",&orr_pct,&orrc_pct,&dcr_pct); If the C, D and E are character variables like COHORT_NEW and N then you need to add quotes around the values.  insert into eff_all2(Cohort_new, N, c, d,e) values ("Total","&ne_allt","&orr_pct","&orrc_pct","&dcr_pct"); Note that if you could have used an expression then the PUT() function would have only worked if C, D and E were CHARACTER variables because formats always generate CHARACTER strings.   What type of numbers are original C, D and E values?  Why did you multiple they mean by 100?  The PERCENT format expects actual percentages as the values.  So 50% would be the number 0.50.  If you multiply by 100 and the display it with the PERCENT format you will get 5000% instead of 50%.   Finally why the heck are you generating the macro variables at all? proc sql; insert into eff_all2(Cohort_new, N, c, d,e) select "Total","&ne_allt",mean(c),mean(d),mean(e) from Efficacy ; quit;   ... View more

This does not make sense: data_null_; length ta compd study $ 10; ta=substr("&study_list",1,2); compd=substr("&study_list",3,3); study=substr("&study_list",7,3); run; %let ccvpath="/gbs/prod/clin/data/&ta./&compd./&study./stable/blinded/level0"; You create three data step variables, which immediately vanish once the data step terminates. And then you try to use three macro variables with the same names, which are never defined anywhere in your code. You probably have messages about undefined macro variables in your log. The macro language provides the %SUBSTR function, if you need to define macro variables. ... View more

@Kp1234 wrote: Thank you for your reply,  I'm using the proc import to read a text file which contains the variable    PROC IMPORT OUT=manifest datafile="../../sd/accumulative/manifests/manifestcomb.txt" REPLACE ; delimiter='09'x; guessingrows=32767; RUN; quit;   when sas reads the variable "Study Code"n (This is one variable), it takes out space and the leading 0's Eg:  0156 00062 becomes: 15600062, is there something that i need to do in the PROC import to keep the formatting?   Using PROC IMPORT will introduce some inconsistency in how the variables are defined since it can only guess what you want.  But it will NOT remove an embedded spaces.  Sounds like your text file does not have spaces between the two parts of the field. And since that makes it look like a numeric value PROC IMPORT will make a numeric variable instead of a character variable.   The real fix is to just write your own program to read the text file instead of using PROC IMPORT.  That will also allow you to fix the goofy variable name with spaces in the middle of it that was used in the code in the original question.   If you already have a numeric variable and you want it to print using nine digits then use the Z9. format.  format subject_code z9.; Or use the Z9. format with a PUT() function call to make a character variable. subject_code_string=put(subject_code,z9.); If you have read actual read the two separate values as individual numeric variables and want to generate the 9 digit string then use the PUT() function on each number separately. subject_code=put(site,z4.)||put(subj,Z5.); ... View more