[jira] [Updated] (LUCENE-7102) LatLonPoint newDistanceSort fails with "sort missing first"

["Robert Muir (JIRA)" <ji...@apache.org>]

     [ https://issues.apache.org/jira/browse/LUCENE-7102?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]

Robert Muir updated LUCENE-7102:
--------------------------------
    Attachment: LUCENE-7102.patch

Here is a patch (with tests for missing, and different missingValues).

First we restrict missingValue to either Double.POSITIVE_INFINITY (missing values last) and Double.NEGATIVE_INFINITY (missing values first). This keeps things simpler and will allow for more optimizations.

In either case we can just bound Integer.MIN_VALUE .. Integer.MAX_VALUE for each dimension:
* +Inf means we've filled the priority queue with only missing values, so any possible value competes.
* -Inf means at this point, only missing values can possibly compete anymore (and only if you have another comparator). I don't think we should do anything tricky to optimize this case.

> LatLonPoint newDistanceSort fails with "sort missing first"
> -----------------------------------------------------------
>
>                 Key: LUCENE-7102
>                 URL: https://issues.apache.org/jira/browse/LUCENE-7102
>             Project: Lucene - Core
>          Issue Type: Bug
>            Reporter: Robert Muir
>         Attachments: LUCENE-7102.patch
>
>
> The distance sort comparator creates bounding boxes when the priority queue is full, to speed up sorting.
> But with missing values (which we don't test), they can be e.g. -Inf/+Inf and we do the wrong thing (e.g. try to make illegal infinite bounding boxes).



--
This message was sent by Atlassian JIRA
(v6.3.4#6332)

---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@lucene.apache.org
For additional commands, e-mail: dev-help@lucene.apache.org