Another common way is to simply group areas in grids and tag each grid with a label that is inserted in a fulltext search.

In general, this is a logical step, since real locations are always part of administrative divisions, like counties, cities, and states. For this reason, it’s good to do a fulltext search to trim down the number of documents for which geo-distance is calculated. If the search does fulltext matching, all other WHERE clauses and expressions declared in the SELECT statement are calculated over the result set provided by the fulltext match. In fullscan searches (no fulltext matching) it assumes calculating a geo-distance for all documents in the index. Here, we do a radius search around point (0.659298124, -2.136602399) in the limit of 10000 meters (10km).Ĭalculating geo-distances is a somewhat expensive operation. Mysql > SELECT *, GEODIST ( 0.659298124, - 2.136602399, latitude, longitude ) as distance FROM geodemo WHERE distance < 10000 ORDER BY distance ASC LIMIT 0, 100 At output, GEODIST() offer the distance in meters between the two points. For example, MySQL has RADIANS() and DEGREES() functions and PHP has deg2rad() and rad2deg(). Most databases and programming languages offer functions to convert from degrees to radians and vice-versa. In general, geo-coordinates are stored as degrees, so you need to convert them at both indexing time and at searching time. At input, the GEODIST() function expects 2 pairs of coordinates in radians – that means float type attributes in Sphinx. Sphinx is able to provide this through the SphinxQL function GEODIST() and it’s API counterpart SetGeoAnchor() which calculates distances between 2 coordinates, taking the earth’s curvature into account.

If you need to provide this kind of search, Sphinx has you covered. They may be expecting to see results within a certain distance from their current location, or they may want to limit their search to a specific region (independent of their current location). When users come to your site and search for, lets say real estate, they will most likely be expecting to have some control over the region in which they are searching. Geo-distance search is a necessary component of any search solution.

Geodist map