When generating an MDM document from a .sav file or accessing the case data with no metadata, the IBM SPSS Statistics SAV DSC handles IBM SPSS Statistics system-missing and user-missing values as described below.

The IBM SPSS Statistics SAV DSC presents IBM SPSS Statistics system-missing values as null values in all MDM variable types except for those based on IBM SPSS Statistics multiple category sets, in which it presents system-missing values as empty category lists. For more information, see "When the Response Set Contains System-Missing Values" in Variable definitions when reading from a .sav file.

A .sav file does not provide enough information for the IBM SPSS Statistics SAV DSC to distinguish between empty and missing category lists.

In IBM SPSS Statistics you can define certain values of a variable as user-missing. For example, for an IBM SPSS Statistics numeric variable, you can define up to three discrete values, a range of values, or a range of values and a single discrete value as user-missing. The way that the IBM SPSS Statistics SAV DSC represents user-missing values depends on how it interprets the IBM SPSS Statistics variable:

For more information about how the IBM SPSS Statistics SAV DSC maps IBM SPSS Statistics variable types to MDM variable types, see Variable definitions when reading from a .sav file.