The following Question & Answer List is based on SAS questions received by the author.
SAS Software Performance Guidelines
When using DB2 to store data, use Pass-Through SQL to read data from DB2 tables into SAS Software, instead of SAS Access Descriptors and Views. Pass-Through SQL allows DB2 to optimise the queries to use DB2 indexes.
The volume of data passed from DB2 to SAS Software has a significant impact on efficiency, so reducing the number of columns and rows passed to SAS Software will reduce the data volume, and hence will improve the performance.
Once the data is stored in SAS datasets, reduce the number of dataset reads by performing more data manipulation in a single step, where appropriate. It is also possible to create many SAS datasets in a single data step, so that a single pass of the input dataset can generate multiple output datasets.