Difference between revisions of "Indexes"

From BaseX Documentation
Jump to navigation Jump to search
Line 58: Line 58:

Revision as of 13:52, 10 December 2010

Existing Indexes

Indexes can speedup queries by magnitudes. Currently, four indexes exist:

  • Text Index: This index speeds up text comparisons in predicates.
  • Attribute Index: This index speeds up attribute value comparisons in predicates.
  • Full-Text Index: Full-text queries are sped up by this index.
  • Path Summary: This index speeds up the resolution of location paths.

Examples of using the indexes

Here are some examples for queries which are rewritten for index access:

Text-Based Queries:

  • //node()[text() = 'Usability']
  • //div[p = 'Usability' or p = 'Testing']
  • path/to/relevant[text() = 'Usability Testing']/and/so/on

Attribute Index:

  • //node()[@align = 'right']
  • descendant::elem[@id = '1']
  • range/query[@id >= 1 and @id <= 5]

Full-Text Index:

  • //node[text() contains text 'Usability']
  • //node[text() contains text 'Usebiliti' using fuzzy]
  • //book[chapter contains text ('web' ftor 'WWW' using no stemming) ftand 'diversity' using stemming distance at most 5 words]

The full-text index is optimized to support all features of the XQuery Full Text Recommendation.

BaseX extends the specification by offering a fuzzy match option. Fuzzy search is based on the Levenshtein algorithm; the longer query terms are, the more errors will be tolerated.

Default "Case Sensitivity", "Stemming" and "Diacritics" options will be considered in the index creation. Consequently, all queries will be sped up which use the default index options.

Index data structures

  • Text/Attribute Index
    Both the text and attribute index are based on a balanced B-Tree and support exact matches and range queries.
  • Full-Text Index (Standard)
    The standard full-text index is implemented as sorted array structure. It is optimized for simple and fuzzy searches.
  • Full-Text Index (Wildcards enabled)
    A second full-text index is implemented as a compressed trie. Its needs slightly more memory than the standard full-text index, but it supports more features, such as full wildcard search.