[
https://issues.apache.org/jira/browse/LUCENE-8634?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16744162#comment-16744162
]
Ignacio Vera commented on LUCENE-8634:
--------------------------------------
{quote}Are there concerns regarding the fact that a polygon might become
invalid through encoding (and eg. cross itself).
{quote}
The encoding is very fine grid with cells of size 8.39e-8 degrees X 8.39e-8
degrees. So when moving to the encoding space it can happen that two points
become the same point or if close enough to an edge in theory it can cross it.
This is the same case as in the tessellator and never happen during the test. I
check and Polygon2D can handle when having shared vertex and could not really
reproduce the other case, so I am not too worry about this.
{quote}Should we keep lines and polygons in the encoded space like boxes?
{quote}
I have been wondering that all the time (It might help performance) but it is a
bigger change. Sure if we move query shapes to the encoding space, then we
should probably keep them there.
> LatLonShape: Query with the same polygon that is indexed might not match
> ------------------------------------------------------------------------
>
> Key: LUCENE-8634
> URL: https://issues.apache.org/jira/browse/LUCENE-8634
> Project: Lucene - Core
> Issue Type: Bug
> Components: modules/sandbox
> Affects Versions: 8.0, 7.7, master (9.0)
> Reporter: Ignacio Vera
> Priority: Major
> Attachments: LUCENE-8634.patch, LUCENE-8634.patch
>
>
> If a polygon with a degenerated dimension is indexed and then an intersect
> query is performed with the same polygon, it might result in an empty result.
> For example this polygon with degenerated longitude:
> POLYGON((1.401298464324817E-45 22.0, 1.401298464324817E-45 69.0,
> 4.8202184588118395E-40 69.0, 4.8202184588118395E-40 22.0,
> 1.401298464324817E-45 22.0))
>
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
