WHERE Clause Planner for Hadoop

The SPD Server WHERE clause planner does the tuning work for you by automatically costing the different approaches to index evaluation. The SPD Server WHERE clause planner avoids computation-intensive operations and uses simple computations where possible.
WHERE clause EVAL strategies SPD Server indexing keeps track of the cardinality ratio and distribution of variable values in a table and uses them to calculate the cost of a WHERE clause. The WHERE clause planner uses four evaluation strategies to determine the number of rows that will be required to execute a given query.
EVAL 6 emulates the behavior of EVAL 2. EVAL 6 means that the query is a candidate for Hadoop WHERE processing. If the Hadoop WHERE processing fails, EVAL 6 reverts to EVAL 2 operation. EVAL 6 takes true rows as determined by EVAL 1, EVAL 3, or EVAL 4, and then eliminates any rows shown to be false, leaving a table that contains only true rows. EVAL 2 processes all rows of a table when no index evaluation is possible. For example, no index evaluation is possible when an index is not present or when some predecessor function performs an operation that invalidates the index.
For more information about the WHERE clause planner and WHERE clause EVAL strategies, see SAS Scalable Performance Data Server: User’s Guide.