In a DATA step, how are labels assigned to variables?

Labels in a DATA step are used to provide more descriptive names for variables, which can enhance readability and clarity in reports or outputs. When a label is assigned to a variable in a DATA step, it becomes a permanent attribute of that variable within the dataset created by the step. This means that once the dataset is saved, the assigned label remains associated with that variable, even after the DATA step has completed execution.

Using the LABEL statement, you can assign a meaningful description to a variable, which will then be stored as metadata. This attribute can be particularly useful when generating reports or viewing data, as it can convey more information than the simple variable name alone.

For instance, if you have a variable named "sales", you might assign it a label like "Total Sales Amount" to clarify its purpose. This label will persist in the dataset and be displayed in output procedures, making it clear to users what that variable represents.

The other choices do not accurately describe how labels function within DATA steps. Labels are indeed supported, and they are not just temporarily assigned; they remain until the dataset is deleted or altered. Therefore, they are not conditional based on specific tasks but are applied consistently as part of the variable's metadata.