Importance

Loading Importance Indicators...

Abstract

A method and apparatus for creating a statistical representation of a query result that can be performed without executing the underlying query. For a binary-join query, a scan is performed on one of the join tables. A multiplicity value that estimates the number of tuples in the other join table that has a matching join attribute to the scanned tuple is calculated. A number of copies (as determined by the multiplicity value) are placed in a stream of tuples that is sampled to compile the statistical representation of the query result. For acyclic-join generating queries including selections, the above procedure is recursively extended. If multiple statistical representations are sought, scans can be shared. Scan sharing can be optimized using shortest common supersequence techniques.

Loading the Abstract Image...

First Claim