Thanks David. I'm going to continue to play with this, as an FYI you were spot on, changing to use a geohash field worked with the previous test. Again I appreciate all of the information, and awesome work.
On Fri, Jul 22, 2011 at 10:05 AM, Smiley, David W. <dsmi...@mitre.org> wrote: > Wrapping the dateline or being able to encircle one of the poles (but not > necessarily both) are polygon query features that I feel need to be addressed > before this module is first released (whenever that is), definitely. And > arguably before benchmarking, which we're looking to focus on soon. So > "Soon". W/i 2 months? Just a guess. > > ~ David > > On Jul 22, 2011, at 8:18 AM, Jamie Johnson wrote: > >> Ah, my mistake then. I will switch to using the geohash field. When >> doing my query I did run it against geohash but when I got Russia that >> was more incorrect than point so I stopped using it. >> >> Is there a timeline by which you expect the dateline issue to be >> addressed? I don't believe that will be an issue for me since I don't >> believe we have any information which will be from that area of the >> world but I'll have to verify. >> >> >> On Fri, Jul 22, 2011 at 3:35 AM, David Smiley (@MITRE.org) >> <dsmi...@mitre.org> wrote: >>> Jamie, >>> You are using the field named "point" which is based on PointFieldType. >>> Keep in mind that just because this field type is named this way, does *not* >>> mean at all that other fields don't hold points, or that this one is >>> especially suited to it. Arguably this one is named poorly. This field >>> type is loosely based on Solr's LatLonType but it doesn't yet fully >>> implement capabilities like crossing the dateline and shapes other than a >>> bounding box. I've enhanced the code to log a warning if the query shape is >>> expanded to the bounding box. Any way, it's a currently a 2nd class citizen >>> in LSP. >>> I recommend you use the field named "geohash" which uses a field type name >>> starting with "Recursive...". Note that the same query you did with this >>> field returned the polygon of Russia. There is an outstanding box in which >>> we don't properly handle polygons crossing the dateline; other shapes are >>> fine. >>> >>> ~ David >>> >>> ----- >>> Author: https://www.packtpub.com/solr-1-4-enterprise-search-server/book >>> -- >>> View this message in context: >>> http://lucene.472066.n3.nabble.com/Geospatial-queries-in-Solr-tp3183374p3190762.html >>> Sent from the Solr - User mailing list archive at Nabble.com. >>> > >
