An index is a separate file from a data set that contains information about observations within the data set. Specifically, an index contains value/identifier pairs that indicate the location of observations within the data set and the value of one or more key variables in that observation.

To create an index at the same time that you create a data set, you use the INDEX= option in the DATA statement. You must assign a unique name to a composite index, while a simple index is automatically assigned the name of the key variable as its name. You can set the value of the MSGLEVEL= system option to I in order to see messages about indexes in the SAS log.

For many maintenance tasks that you perform on a data set, SAS automatically performs corresponding tasks to the index file. For example, if you delete a data set, the index file is deleted as well. If you rename a data set with the CHANGE statement in the DATASETS procedure, SAS automatically renames the index file. If you copy a data set to a new location with the COPY statement in the DATASETS procedure, SAS automatically reconstructs the index file in the new location.

You can use the DATASETS procedure or the SQL procedure to create or delete an index from an existing data set. You can also rebuild the index with a DATA step and use the INDEX= option to create an index on the rebuilt data set. However, rebuilding a data set uses more system resources than adding an index to an existing data set with either the DATASETS or the SQL procedure.

You use the CREATE INDEX statement of the SQL procedure to create an index on an existing data set. In the SQL procedure, you must name the index in the CREATE INDEX statement; for a simple index, the index name must match the name of the key variable.

You can use either the CONTENTS procedure or the CONTENTS statement in the DATASETS procedure to generate a list of information about a data set, including a list of existing indexes. All indexes for a data set are stored in a single file that is separate from but has the same name as the data set.

An index can improve the efficiency with which SAS is able to access certain observations in a data set. However, an index is not always useful. SAS will not use an index to process subsetting IF statements, or other statements that SAS determines might be more efficiently processed without an index.