System and method to enable parallel text search using in-charge index ranges

A document retrieval system for searching a document database for documents stored in the document database matching a specified query condition, comprising:

• (1) a storage device, arranged to:
• (1)-1:
• store information for identifying each of a plurality of search servers in the document retrieval system;
• (1)-2:
• store a character string included in the documents with an association to:
• a document identifier of a document including the character string;
• and
• a page identifier of a page on which the character string is indexed;
• and
• (1)-3:
• store a document range with an association to the character string and the page identifier on which the character string is indexed;
• (2) a system management server, arranged to:
• (2)-1:
• receive the specified query condition from the storage device, and determine a range of documents to be searched by each of the search servers according to:
• the specified query condition, and
• the document identifier and the page identifier stored in the storage device,
• (2)-2:
• send information on:
• the range of documents determined to be searched to the storage device at which the range of documents to be searched is recorded, and
• the specified query condition to each of the search servers at which the specified query condition is recorded;
• (3) wherein each of the plurality of search servers is arranged to:
• (3)-1:
• acquire a document identifier, and acquire a page identifier based on the document identifier, for each of the plurality of search servers, and
• (3)-2:
• perform a search operation of the document database at each of the search servers based on the specified query condition and the acquired page identifier;
• and
• (4) wherein the system management server is further arranged to:
• (4)-1:
• output a total search operation result derived from the search operations performed for all of the plurality of search servers;
• wherein the system management server is further arranged to:
• compare processing costs for indexes when the indexes for two or more character strings are referred to execute the searching operation;
• and
• determine the range of documents set for the character string having the larger processing cost as range of documents for all the character strings as a result of said comparison.

A computer-readable storage medium storing a document retrieval/execution program which, when executed by a computer, causes the computer to search a document database for documents stored in the document database matching a specified query condition using a system management server, a plurality of search servers, and a storage device, said program comprising:

• acquire a document identifier of a document including a character string and a page identifier of a page on which the character string is indexed, based on the document identifier, for one of the plurality of search servers;
• perform a search operation of the document database at the search server based on the specified query condition and the acquired page identifier;
• compare processing costs for indexes when the indexes for two or more character strings are referred to execute the search operation;
• determine a range of documents to be searched, set for the character string having the larger processing cost, as the range of documents to be searched for all character strings as a result of said comparison;
• and
• send the result of the search operation to the system management server for aggregation with search results from each of the other search servers and output by the system management server.