You don't really describe your issue, but perhaps the most likely is that PROC IMPORT didn't look at enought observations to make a good guess at how long the character fields should be.
You could try adding a GUESSINGROWS statement.
PROC IMPORT DATAFILE="C:\Have.txt"
OUT=Want replace
DBMS=dlm
;
delimiter=',';
GETNAMES=YES;
guessingrows=32000;
RUN;
But to really read a CSV file just write the data step yourself. Since it is your data you probably had a better idea of how you want to read it than SAS can guess. You could just make up dummy variable names.
data want ;
infile 'have.txt' dsd truncover firstobs=2 ;
length var1-var10 $200 ;
input var1-var10;
run;
Or you could copy the first row of the file and use it to make the variable names. You can set the type and length in a LENGTH statement. Attach any informats that are needed to read the data from the CSV file. Attach any formats that are useful. Avoid using INFORMAT or FORMAT statement as a substitute for defing your variables explicitly with a LENGTH or ATTRIB statement. Avoid attaching $xx formats/informats to character variables as they do not add value and can cause trouble when combining datasets.
Then the INPUT statement is trivial.
data want ;
infile 'have.txt' dsd truncover firstobs=2 ;
length id $20 date 8 name $50 ;
informat date mmddyy.;
format date yymmdd10.;
input id -- name ;
run;
