getting error message: variable defined as both character and numeric

[crc7rb]

Hello,

I am merging 22 years of data sets in SAS 9.4 and I am getting several error statements saying variables such as sex, race, zipcode are defined a both character and numeric.

sex is M(male), F(female), U(unknown)

race is c 1(AIAN), 2(A), 3(BAA), 4(NHOPI), 5(W), 6(O), 7(U)

zipcode is continuous numbers

Will I have to rename these variables in each data set before I merge them?

This was my merging code:

data IHD9214;
set Ahca.Ihd1992 Ahca.Ihd1993 Ahca.Ihd1994 Ahca.Ihd1995 Ahca.Ihd1996 Ahca.Ihd1997 Ahca.Ihd1998 Ahca.Ihd1999 Ahca.Ihd2000 Ahca.Ihd2001 Ahca.Ihd2002 Ahca.Ihd2003 Ahca.Ihd2004 Ahca.Ihd2005 Ahca.Ihd2006 Ahca.Ihd2007 Ahca.Ihd2008 Ahca.Ihd2009 Ahca.Ihd2010 Ahca.Ihd2011 Ahca.Ihd2012 Ahca.Ihd2013 Ahca.Ihd2014;
run;

[Kurt_Bremser]

First of all, this is not a merge, it's a simple sequential set.

And yes, you need to have all variables of the same name to be of the same type (same length is not a must, but very advisable!).

So inspect your datasets and make the necessary changes before concatenating them.

[Reeza]

Do all the files have the same variables and types? Did you happen to import these files from an XLSX or CSV/txt file?

[ballardw]

This code may be useful in indicating which variable/datasets you need to be concerned with. It will create a report table with an X indicating which datasets have a variable of which type.

proc format library=work;
   value x
   1='X'
   ;
proc tabulate data=sashelp.vcolumn;
  where libname='AHCA';
  class memname name type;
  tables name*type,
         memname*n=''*f=x.
        /misstext=' ';
run;

If these sets came from a common structure and were read with proc import then you may want to go back to reading the data with a method that will set consistent types and lengths.

Your error is likely to have issues with data truncation do to length differences as well.