You minimize the CPU time that SAS uses when you execute the minimum number of statements in the most efficient order.

For a more efficient program, place the subsetting IF statement as soon as logically possible in a DATA step when creating a subset of data.

Review guidelines for using conditional logic efficiently with IF-THEN/ELSE statements or SELECT statements. Remember to minimize the number of statements that use SAS functions or arithmetic operators.

Conditionally execute only necessary statements by placing statements in groups that are associated with IF-THEN/ELSE statements or SELECT/WHEN statements. Groups of statements execute only when a particular condition is true. Review the criteria for using DO groups efficiently.

You should avoid reading or writing data more than necessary in order to minimize I/O operations.

There are a variety of techniques that you can use. For example, use a single DATA step to create multiple output data sets from one pass of the input data, rather than using multiple DATA steps to process the input data each time you create an output data set. Create sorted subsets by subsetting data with the SORT procedure rather than subsetting data in a DATA step and then sorting. Change variable attributes by using PROC DATASETS rather than a DATA step.

If you process fewer observations and variables, SAS performs fewer I/O operations. To limit the number of observations that are processed, you can use the subsetting IF statement and the WHERE statement. Best programming practices can be applied if you understand other differences between subsetting IF and WHERE statements. You can also improve performance by applying OBS= and FIRSTOBS= processing with a WHERE statement.

To select observations when reading data from external files, position a subsetting IF statement in a DATA step so that it reads only the variables that are needed to select the subset before reading all the data. This can reduce the overhead required to process data.

To limit the number of variables that are processed, you can use either of the following:

In the SET statement, the DROP= or KEEP= data set option controls which variables are read and subsequently processed. In the DATA statement, the DROP= or KEEP= data set option controls which variables are written to an output data set after processing. Using the SET statement with these options is the most efficient and best practice.

When you use SAS to repeatedly analyze or manipulate any particular group of data, create a SAS data set instead of reading the raw data each time.

Reading data from an external file rather than reading from a SAS data set greatly increases CPU usage.

Invoke procedures once rather than multiple times. Several procedures enable you to create multiple reports by invoking the procedure only once.

Using a single DATASETS procedure instead of multiple DATASETS procedures to modify the descriptor portion of a data set results in a noticeable savings in both CPU and I/O operations. Also, you can take advantage of RUN-group processing to submit groups of statements without ending the procedure.