jpountz commented on pull request #749:
URL: https://github.com/apache/lucene/pull/749#issuecomment-1083384716


   Interestingly #687 would further specialize this query for longs by relying 
on the `LongPoint` encoding. cc @wjp719
   
   > I was phasing out FLOAT / DOUBLE support, as per your comment, but the 
implementation we will end up with becomes very inconsistent and confusing, for 
example SortedNumericSortField happily deals with FLOAT / DOUBLE
   
   Since this `IndexSortSortedNumericDocValuesRangeQuery` is in sandbox, it's 
totally fine if it looks a bit half-baked and only works with a small subset of 
the numeric types.
   
   In my opinion, long-term this `IndexSortSortedNumericDocValuesRangeQuery` 
should no longer be exposed to users. Lucene would have long/int/double/float 
fields ([LUCENE-10162](https://issues.apache.org/jira/browse/LUCENE-10162)) 
that would have factory methods to produce queries, and both 
`IndexOrDocValuesQuery` and `IndexSortSortedNumericDocValuesRangeQuery` would 
become implementation details of these queries that are not exposed to users. 
And maybe only long and int fields would optimize for the case when they are 
the primary index sort, but that would be hidden from users and wouldn't affect 
usability.


-- 
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

To unsubscribe, e-mail: [email protected]

For queries about this service, please contact Infrastructure at:
[email protected]



---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]

Reply via email to