Browse Prior Art Database

Optimization of Text Search System Combined with Database

IP.com Disclosure Number: IPCOM000012250D
Original Publication Date: 2003-Apr-23
Included in the Prior Art Database: 2003-Apr-23
Document File: 2 page(s) / 19K

Publishing Venue

IBM

Abstract

This is the disclosure to optimize the combined search for both database and text search system.

This text was extracted from a PDF file.
At least one non-text object (such as an image or picture) has been suppressed.
This is the abbreviated version, containing approximately 55% of the total text.

Page 1 of 2

THIS COPY WAS MADE FROM AN INTERNAL IBM DOCUMENT AND NOT FROM THE PUBLISHED BOOK

JP820020879 Koichiro Kato/Japan/IBM Junichi Satoh, Kohichi Toshioka

Optimization of Text Search System Combined with Database

The current text search and database integrated system outputs the search  result by merging both text search results and database retrieval results.  Also, it merges these results after the independent operation, which may  have the unnecessary work if the one result set has very large data  and  the other does not for the process of the final combined result. 

This disclosure deploys the selectivity parameters to estimate the number  of hits and optimizes the sequence for the combined operation.  This method checks each query term's selectivity parameter for both  database and search engine.  If the hit ratio of one side is smaller than  the other side significantly, the smaller result set should be executed  first. Then the result  data should be transferred to the other side. This  method eliminates the extra work and reduce the transmission time between  two subsystems, which  results in the higher throughput.                   

Fig 1 (typical case)in the attached chart shows the outline of the current  combined search system.  In this case, text search has much smaller result set.      The "AND" operation needs the extra large efforts to combine both data if  one side is significantly smaller than the other side.   

Fig 2 shows the proposed combined system which deploys the selectivity  comparison system.  1,2 and 3,4 checks the selectivity parameters and  the search planning  engine 13 optimizes the execution order according to the selectivity  information.  In this case the load of database and "AND" operation is smaller than  previous case for the same situation described in Fig1.    (If the text  search generates much larger result, which is rev...