It is more efficient to use indexed
access for a small subset and to use sequential access for a large
subset. Therefore, after identifying any available indexes and evaluating
the conditions in the WHERE expression, SAS estimates the number of
observations that will be qualified by the index. Whether SAS uses
an index depends on the percentage of observations that are qualified
(the size of the subset relative to the size of the data set), as
If the subset is less than 3% of the data set, direct access is almost certainly more efficient than sequential access, and SAS will use ...