Updates for structure indexes

  • Authors:

  • Raghav Kaushik;Philip Bohannon;Jeffrey F. Naughton;Pradeep Shenoy

  • Affiliations:

  • Univ. of Wisconsin, Madison;Lucent Technologies, Bell Laboratories;Univ. of Wisconsin, Madison;Univ. of Washington, Seattle

  • Venue:
  • VLDB '02 Proceedings of the 28th international conference on Very Large Data Bases
  • Year:
  • 2002

Quantified Score

Hi-index 0.00

Visualization

Abstract

The problem of indexing path queries in semistructured/XML databases has received considerable attention recently, and several proposals have advocated the use of structure indexes as supporting data structures for this problem. In this paper, we investigate efficient update algorithms for structure indexes. We study two kinds of updates -- the addition of a subgraph, intended to represent the addition of a new file to the database, and the addition of an edge, to represent a small incremental change. We focus on three instances of structure indexes that are based on the notion of graph bisimilarity. We propose algorithms to update the bisimulation partition for both kinds of updates and show how they extend to these indexes. Our experiments on two real world data sets show that our update algorithms are an order of magnitude faster than dropping and rebuilding the index. To the best of our knowledge, no previous work has addressed updates for structure indexes based on graph bisimilarity.