The programmer can create or define formats to provide a more meaningful description to the values of a variable.
Formats can be either temporary (e.g., they only exist within the program) or permanent (e.g., they are saved as a file on the server or computer). Formats for a character variable always start with $, while numeric formats can start with any letter.
A simple data check is looking at formatted frequencies to see how much data are missing and non-missing. The following will simply show the number of missing and non-missing values for all variables in the dataset. The character variables will use the $charf format, while the numeric variables will use the numf format.
Sometimes there are just too many values to list out individually in a PROC FORMAT VALUE statement. In these cases, it is possible to create a format using a dataset. The following example shows how it may be done.
In this example, we are assigning thousands of procedure codes into groups. The researcher has given the programmer his group definitions for specific CPT codes in a file called procedures.txt. Tabs were used to separate the code from the group.