@arnreddy007 wrote:
SAS considering numeric coloumns with missing values as characters while I am doing import with proc import I have tried both in Base SAS and SAS EG, is there a way around?
Show the code, or the code generated if using a wizard. File type has some bearing, import options depending on file type such as GUESSINGROWS can make a difference.
If the "column" ends up with a $1. format, character of length 1, then the import procedure saw missing values for all of the rows used to GUESS which type of variable should be in the variable.
If the file was something like a spreadsheet you may have better luck doing a File>Save As to CSV format before attempting Proc import. Set the value for the number of rows to guess in the wizard to something 'large', where large ideally is the number of rows in the file but a value in the 1000's usually works for much larger files.
Note that Proc Import always guesses and depending on the individual file can generate different data types and lengths of character variables from two files that supposedly have the same file layout. Which often causes problems when the resulting data sets are combined.
Click image to register for webinar
