Change variable from character to numeric after XLSx import

[amigapomba]

Hi all,

I'm imported an XLSx file into SAS University edition. For some reason three of the numeric variables have been classified by SAS as character variables, even though no observations include any characters. Can anyone tell me how to correct this?

Here is my import code:

PROC IMPORT DATAFILE="/folders/myfolders/Med.xlsx"
OUT=WORK.med2
DBMS=XLSX
REPLACE;

TIA for any help!

[Reeza · Accepted solution]

Check the type in the Excel file. There isn't really a good way to control the import from an Excel file, the usual workaround suggestions are to create a CSV file instead and import that, it allows you to have full control. It's really rare to have a full column of just numbers be imported as characters though. How is missing denoted in that file?

[learsaas]

try ：

mixed=yes

[amigapomba]

Hi thanks very much for your help-

I looked closer at the entries for those variables in excel and saw that they each have one or two that were somehow entered as text. I will try to change them to numbers and repeat the import. Missing data in the Excel file are blank cells, for those three variables as well as for the seven other variables that sas correctly labeled as numeric.

[amigapomba]

thanks again! after i corrected those cells the import worked and all numeric variables were classified as numeric.