You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
At the moment, we support auto-complete in the Lucene indexes by adding a separate directory with auto-complete specific information. However, that has the disadvantage of requiring duplicate storage of the field and index data for auto-complete purposes. To save space, we should consider using the main Lucene directory as the search index for auto-complete, though this will have the disadvantage that it will require re-loading the original record and then extracting the text from it in order to support auto-complete, but that might still be worth it given the expense of these indexes.
The text was updated successfully, but these errors were encountered:
…ucene directory
This modifies the auto-complete cursor so that it reads from the main Lucene index and then reconstitutes the original text by reading from the base record data.
alecgrieser
added a commit
to alecgrieser/fdb-record-layer
that referenced
this issue
May 19, 2022
…ucene directory
This modifies the auto-complete cursor so that it reads from the main Lucene index and then reconstitutes the original text by reading from the base record data.
…ory (#1683)
* Current State is rough
* Resolves#1682: Support auto-complete via querying main Lucene directory
This modifies the auto-complete cursor so that it reads from the main Lucene index and then reconstitutes the original text by reading from the base record data.
Co-authored-by: john_leach <jleach4@gmail.com>
At the moment, we support auto-complete in the Lucene indexes by adding a separate directory with auto-complete specific information. However, that has the disadvantage of requiring duplicate storage of the field and index data for auto-complete purposes. To save space, we should consider using the main Lucene directory as the search index for auto-complete, though this will have the disadvantage that it will require re-loading the original record and then extracting the text from it in order to support auto-complete, but that might still be worth it given the expense of these indexes.
The text was updated successfully, but these errors were encountered: