Login

Find nearby objects

Author:
paulsmith
Posted:
April 18, 2007
Language:
Python
Version:
.96
Tags:
distance gis postgis geography geometry nearby nearest
Score:
2 (after 2 ratings)

This code assumes a) that you are using PostgreSQL with PostGIS and b) that the geometry column in your model's table is populated with points, no other type of geometry.

It also returns a list instead of a QuerySet, because it was simpler to sort by distance from the given point and return that distance as part of the payload than to muck around with QuerySet. So bear in mind that any additional filtering, excluding, &c., is outside the scope of this code.

'Distance' is in the units of whatever spatial reference system you define, however if you do not intervene degrees decimal is used by default (SRID 4326, to be exact).

To get distance in units a little easier to work with, use a spatial ref close to your domain set: in my case, SRID 26971 defines a projection centered around Illinois, in meters. YMMV.

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
# in your models.py or managers.py

class PlaceManager(models.Manager):
    def nearest_to(self, point, number=5, geom_column='geom', from_srid=4326, to_srid=4326):
        """finds the model objects nearest to a given point
        
        `point` is a tuple of (x, y) in the spatial ref sys given by `from_srid`
        
        returns a list of tuples, sorted by increasing distance from the given `point`. 
        each tuple being (model_object, dist), where distance is in the units of the 
        spatial ref sys given by `to_srid`"""
        if not isinstance(point, tuple):
            raise TypeError
        from string import Template
        cursor = connection.cursor()
        x, y = point
        table = self.model._meta.db_table
        sql = Template("""
            SELECT id,
                Length(Transform(SetSRID(MakeLine($geom_column, GeomFromText('POINT(' || $x || ' ' || $y || ')', $from_srid)), $from_srid), $to_srid))
            FROM $table
            WHERE $geom_column IS NOT NULL
            ORDER BY Distance(GeomFromText('POINT($x $y)', $from_srid), $geom_column) ASC
            LIMIT $number;
        """)
        cursor.execute(sql.substitute(locals()))
        nearbys = cursor.fetchall()
        # get a list of primary keys of the nearby model objects
        ids = [p[0] for p in nearbys]
        # get a list of distances from the model objects
        dists = [p[1] for p in nearbys]
        places = self.filter(id__in=ids)
        # the QuerySet comes back in an undefined order; let's
        # order it by distance from the given point
        def order_by(objects, listing, name):
            """a convenience method that takes a list of objects,
            and orders them by comparing an attribute given by `name`
            to a sorted listing of values of the same length."""
            sorted = []
            for i in listing:
                for obj in objects:
                    if getattr(obj, name) == i:
                        sorted.append(obj)
            return sorted
        return zip(order_by(places, ids, 'id'), dists)

# ex. REPL

>>> Places.objects.nearest_to((-87.96, 42.06), number=5, to_srid=26971)
[(<Place: 600 S See Gwun Ave, Mount Prospect, IL 60056, USA>, 1908.34541121444), (<Place: 48 Raupp Blvd, Buffalo Grove, IL 60089, USA>, 9209.1395558741897), (<Place: 1000 S Milwaukee Ave, IL 60090, USA>, 7996.2420822250497), (<Place: 1175 Howard Ave, Des Plaines, IL 60018, USA>, 8095.9994070327703), (<Place: 701 Thorndale Ave, Wood Dale, IL 60191, USA>, 9613.4684521103609)]

More like this

  1. GeoDjango maps in admin TabularInlines by alanB 4 years, 6 months ago
  2. Updated: GeoJSON Serializer for GeoDjango (gis) by danielsokolowski 3 years, 5 months ago
  3. Easier chainability with custom QuerySets by bendavis78 3 years, 1 month ago
  4. Lorem Ipsum - Random content for your mockups by gfranxman 8 years, 2 months ago
  5. Export queryset to Excel workbook by masida 7 months, 2 weeks ago

Comments

Please login first before commenting.