One easy diagnostic for PROC Import and delimited files is to read the log immediately after running proc import. If there is an error the SAS error messages are usually pretty good about pointing out where things go wrong. If the data result looks unexpected but there is no error the log will show the Datastep code SAS generated to read the file. Often type mismatches (you expect numeric but get character or vice versa) will be shown by reading the INFORMAT statements. You can actually copy that code from the log, make changes and rerun.   Note that Proc Import makes guesses about your data. The Guessingrows statement tells how many rows to look at before making those guesses. If a variable is blank for the first few rows, I think 20 is default, then the variable will be character. If it appears to be numeric in the first few rows but latter has character values such can happen with many code type values such as product identifiers, accounts, then the later values will be missing as they are not numeric. A change of Informat from BEST12. or similar to $12. would correct the issue.   Also the variable names will appear in the code as well as notes about column headings that may not have been acceptable.   So to reiterate: when using proc import read the log before going any further with the project. Addressing issues at this step may save LOTS of headaches later. ... View more