[postgis-users] <-> operator does not return true distance

[Brian M Hamlin]

 Hi Jakob -
 
  I am interested in this topic, but have not looked into it for a while.. 
I wrote a blog post long ago which may explain the situation, as it was 
at that time.. [0]
 
  http://blog.light42.com/wordpress/?s=knn
 
 
  best regards
    --Brian

 
--
(I asked a similar question here 
https://gis.stackexchange.com/questions/247034/postgis-operator-does-not-return-true-distance 
)

I used the <-> operator with PostGIS 2.2.2 and PostgreSQL 9.5.7 . 

According to the docs 
(http://postgis.net/docs/manual-2.2/geometry_distance_knn.html) the <-> 
operator should order by "true KNN distance" and *not* by "centroid 
distance". 

However it does not work for me. See the minimal example below:
( illustration: https://i.stack.imgur.com/QZA6D.png )

SELECT id,
ST_Distance(my_point.geom,
my_polygons.geom) AS real_distance,
ST_Distance(my_point.geom,
ST_Centroid(my_polygons.geom)) AS centroid_distance

FROM (
SELECT 'blue' AS id, ST_Polygon(ST_GeomFromText('LINESTRING(0 3, 2 3, 2 
4, 0 4 ,0 3)'),4326) AS geom
UNION
SELECT 'green' AS id, ST_Polygon(ST_GeomFromText('LINESTRING(2 0, 8 0 , 
8 2, 2 2 ,2 0)'),4326) AS geom

) AS my_polygons,

(
SELECT ST_SetSRID(ST_Point(1, 1),4326) AS geom
) AS my_point

ORDER BY my_point.geom <-> my_polygons.geom

--- obtained Result:

| | real_distance | centroid_distance |
---------------------------------------------
| blue | 2 | 2.5 |
| green | 1 | 4 |

(I expected the rows to be in the opposite order)

Apparently the <-> operator orders by the centroid_distance and not by 
the real_distance, even though the docs say: "[...] for PostgreSQL 9.5 
, does true KNN distance search giving true distance between geometries 
[...]"

I tried the same example on 2.3.2 and PostgreSQL 9.5.7 and I get the 
expected result (which orders by "real_distance"). 

So, for me it seems that ordering by "real_distance" with the <-> 
operator does not work with PostGIS 2.2.2 - but this is *not* written 
in the docs. 

Can you reproduce this behaviour? Or did I understand/make something 
wrong here?

Thanks and best regards,
Jakob

--
Brian M Hamlin
OSGeo California Chapter
blog.light42.com