When a non empty crossjoin returns an empty set, the SAS OLAP Server cannot optimize the query because optimization might result in incorrect or invalid data. Because the query is not optimized, performance is negatively impacted under certain conditions. For example, results will return more slowly if the dimensions that are included in the crossjoin are of high cardinality (significantly greater than 100 unique values), or if there are a large number of calculated measures on the opposite axis from the crossjoin.

Limiting the number of calculated measures in a single query or designing a cube that avoids direct access to high cardinality levels should help to avoid this behavior.

Operating System and Release Information

A fix for this issue for Threaded Kernel OLAP Server Component 9.3_M1 is available at:

A fix for this issue for Base SAS 9.21_M3 is available at:

A fix for this issue for Base SAS 9.3_M1 is available at: