[OAK-9155] oak-search-elastic: remove :fulltext field

• Type: Task
• Status: Open
• Priority: Major
• Resolution: Unresolved
• Component/s: elastic-search, oak-search
• Assignee: Fabrizio Fortino
• Reporter: Fabrizio Fortino
• Created: 26/Jul/20 08:53
• Updated: 26/Jul/20 08:53

Description

While executing the full-text performance benchmarks, where we import a Wikipedia dump into the repository, I have noticed we copy the value of analyzed/nodeScopeIndexed fields into the :fulltext field. AFAIK this is mainly used for full-text queries with a clause like:

contains(., "some text")

The proposal is to remove the :fulltext field completely and for queries clauses like above we can use a multi-match query with cross_fields type [0].

This will impact in a good way the configuration: in lucene a reindex is needed when a property is set to nodeScopeIndex=true. In elastic the reindex won't be needed because the fields for the multi-match query are dynamic.

[0] https://www.elastic.co/guide/en/elasticsearch/reference/current/query-dsl-multi-match-query.html#type-cross-fields