[
https://issues.apache.org/jira/browse/LUCENE-7179?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15226848#comment-15226848
]
Robert Muir commented on LUCENE-7179:
-------------------------------------
{quote}
Its needed for GeoPointField.
{quote}
Not true: we can take float instead. And maybe this is a good idea if we are
not going to handle this stuff properly. Then no data from the user is lost: 64
bits in, 64 bits out.
But today it is 128 bits in, 64 bits out. My problem is exactly how this
lossiness occurs. It must be something we can reason about. For example with
LatLonPoint, for any 64-bit double from the user, we know we encode an integer
that that is always <= to that value (not willy nilly rounding), we know that
sort order is completely preserved in integer space, and that there is no
overflow. This is really all I am asking for.
> GeoPoint and LatLonPoint test data should quantize once
> -------------------------------------------------------
>
> Key: LUCENE-7179
> URL: https://issues.apache.org/jira/browse/LUCENE-7179
> Project: Lucene - Core
> Issue Type: Bug
> Reporter: Nicholas Knize
> Attachments: LUCENE-7179.patch
>
>
> {{LatLonPoint}} and {{GeoPointField}} tests pre quantizes test data to ensure
> consistency with indexed (encoded) data. The pre quantized data then becomes
> indexed, undergoing another quantization. To guarantee numerical stability
> this should be changed such that the test data is quantized after indexing.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
