Importance

Loading Importance Indicators...

Abstract

A query processing system for processing queries in connection with a document text base which has entries each identifying a document and a word in the document. The query processing system includes a plurality of processing elements for processing data in response to commands, and a control arrangement for controlling the processing elements in parallel. The control arrangement first enables the processing elements to generate a segmented posting file having entries, at least some of which have a word identifier and a document identifier. The entries form an array all of whose entries with the same document identifier are contained within one column. The rows of the segmented posting file are aggregated into segments each having a selected number of rows with each segment containing entries having word identifiers within an identified word identifier range. Thereafter, the control arrangement enables the processing elements to use the segmented posting file to process, in parallel, a query in a series of iterations each with respect to a query word. In each iteration, the processing elements receive respective portions of columns comprising a segment of the segmented posting file associated with the word identifier range containing the query word, then identify entries in the segment whose word identifiers correspond to the query word, and finally modify a score maintained for the document identified in the identified entry. Those documents which have a selected score at the end of the series of iterations have the required relationship to the query.

Loading the Abstract Image...

First Claim