Fork me on GitHub
Oak logo
Built by Maven

Elastic Index

Oak supports Elasticsearch (Elastic for short) based indexes for both property constraint and full text constraints. Elastic indexes support similar features as Lucene indexes, however there are differences:

  • The type is elasticsearch.
  • The index definition needs to be under /oak:index. Other locations are not supported.
  • The async property needs to be set to elastic-async. Synchronous, nrt or other lanes are not supported. Indexes are updated asynchronously.
  • refresh is ignored. Changes take effect immediately after changing them. Existing documents in Elasticsearch are not changed.
  • Indexes are NOT automatically built when needed: They can be built by setting the reindex property to true or by using the oak-run tool. We recommend to build them using the oak-run tool.
  • evaluatePathRestrictions cannot be disabled. The parent paths are always indexed. Queries with path restrictions are evaluated at index level when possible, otherwise they are evaluated at repository level.
  • codec is ignored.
  • compatVersion is ignored.
  • useIfExists is ignored.
  • blobSize is ignored.
  • name is ignored.
  • indexPath is ignored.
  • analyzers support the Lucene configuration plus Elasticsearch specific options. Since Elasticsearch uses a more recent version of Lucene compared to the one in oak-lucene module, there might be differences in configuration options that could require changes when migrating from Lucene to Elasticsearch. The HunspellStem filter is not supported since dictionary files are required in the Elasticsearch cluster filesystem.
  • useInExcerpt does not support regexp relative properties.
  • For property definitions, sync and unique are ignored. Synchronous indexing, and enforcing uniqueness constraints is not currently supported in elastic indexes.
  • The behavior of dynamicBoost differs slightly between Lucene and Elasticsearch:
    • Lucene: Boosting is applied at indexing time.
    • Elasticsearch: Boosting is applied at query time.

Full-text queries automatically use dynamically boosted values to match relevant results, but this behavior may not always be desirable. To use these values exclusively for influencing relevance without affecting matching, configure the property definition as follows:

{
  "dynamicBoost": true,
  "useInFullTextQuery": false
}
  • The behavior of suggest is slightly different: For Lucene indexes, the suggestor is updated every 10 minutes by default and the frequency can be changed by suggestUpdateFrequencyMinutes property in suggestion node under the index definition node. In Elastic indexes, there is no such delay and thus no need for the above config property. This is an improvement in ES over lucene.