Lost Variables when Importing dBase, Excel or other Data

Some versions of dBase, Excel and other software have a built-in limitation of the number of variables that can be moved from those packages into SPSS or SAS. One known dBase limit is 255 variables, but this number may vary from software to software, or even from version to version in a particular software package. This is not an SPSS or SAS problem, but a problem with the software that originally created the data file.

Symptoms of this problem include the fact that some variables (columns) get imported into SPSS or SAS correctly, and the data look fine, but not all variables make the transition. This has been noticed with .DBF type files, but may occur with other types as well. In general, this problem is less and less frequent as the other software packages become more sophisticated.

The solution to this problem that is most often recommended is to split the data file into less-than-255 variable portions, making sure to assign an unique id-number variable to each of the split portions, then move the data into SPSS or SAS, and then merge the portions back together.