Pagewise preview ]

CategoryValue
Available viahttp://dbpubs.stanford.edu/pub/2003-13
Submitted on 18th of February 2003
Author Su, Qi; Widom, Jennifer
Title Indexing Relational Database Content Offline for Efficient Keyword-Based Search
Date of publication February 2003
Citation Su, Qi; Widom, Jennifer. Indexing Relational Database Content Offline for Efficient Keyword-Based Search,
Number of pages 13
Language English
Project Database Group
Type Technical Report
Subject group Miscellaneous
Abstract Information Retrieval systems such as web search engines offer convenient keyword-based search interfaces. In contrast, relational database systems require the user to learn SQL and to know the schema of the underlying data even to pose simple searches. We propose an architecture that supports highly efficient keyword-based search over relational databases: A relational database is "crawled" in advance, text-indexing virtual documents that correspond to interconnected database content. At query time, the text index supports keyword-based searches with instantaneous response, identifying database objects corresponding to the virtual documents matching the query. Our system, EKSO, creates virtual documents from joining relational tuples and uses the DB2 Net Search Extender for indexing and keyword-search processing. Experimental results show that index size is manageable, query response time is indeed instantaneous, and database updates (which are propagated incrementally as recomputed virtual documents to the text index) do not significantly hinder query performance. We also present a user study confirming the superiority of keyword-based search over SQL for a wide range of database retrieval tasks.
Keywords keyword search, information retrieval, text database
Contact address qi@db.stanford.edu
Sponsored by National Science Foundation under grant IIS-9817799 and an NSF Graduate Research Fellowship.
Fulltext source
  • Postscript (ps, ps.gz, ps.zip)
  • PDF (pdf, pdf.gz, pdf.zip)
  • Plain text (text, text.gz, text.zip)
  • Management of the document bysiroker@db.stanford.edu

    Pagewise preview ]


    Stanford InfoLab Publication Server