Github user alessandrobenedetti commented on a diff in the pull request:
https://github.com/apache/lucene-solr/pull/398#discussion_r195689810
--- Diff:
lucene/suggest/src/java/org/apache/lucene/search/suggest/InputIterator.java ---
@@ -34,7 +34,7 @@
public interface InputIterator extends BytesRefIterator {
/** A term's weight, higher numbers mean better suggestions. */
--- End diff --
Hi Michael,
The reason to allow for null at the InputIterator level is to distinguish
it from an explicit 0 weight.
In the DocumentDictionary this translates in differentiating when the
weight field was missing for the original document ( NULL ) in opposition to
when the weight field was present and with 0 value.
At this level we just want to ensure that the same behavior is maintained
when we build the auxiliary index :
i.e. if the weight field was missing for the original document, I want it
to be null for the auxiliary index as well.
How the different suggesters implementation will use this to return a
suggestion score, I think will depend on a case by case scenario.
Did I misunderstand anything here ?
---
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
