[NXP-30785] Use elastic word_delimiter_graph instead of word_delimiter to avoid indexing error - Nuxeo Issue Tracker

XML

Word

Printable

Release Notes Summary:
Elastic mapping has been updated to avoid indexing error because of a deprecated word delimiter.
Tags:
- nxplatform
- platform-review
Upgrade notes:

Hide

If you encounter indexing errors because of "negative startOffset", your mapping needs to be updated.
If you have overridden the Elastic mapping then follow the recommended changes in the ticket.
Then you need to reindex the repository, visit https://doc.nuxeo.com/nxdoc/elasticsearch-setup/#rebuilding-the-repository-index for more information.

Show
If you encounter indexing errors because of "negative startOffset", your mapping needs to be updated. If you have overridden the Elastic mapping then follow the recommended changes in the ticket. Then you need to reindex the repository, visit https://doc.nuxeo.com/nxdoc/elasticsearch-setup/#rebuilding-the-repository-index for more information.
Team:
PLATFORM
Sprint:
nxplatform #59
Story Points:
3

The word_delimiter filter can produce indexing errors like:

startOffset must be non-negative, and endOffset must be >= startOffset, and offsets must not go backwards

Since Lucene 7.x the IndexWriter more aggressive about rejecting backwards offsets: https://issues.apache.org/jira/browse/LUCENE-7626.

To fix this, you should be able to replace word_delimiter with word_delimiter_graph, which has offset correction logic.