Selection¶

Attribute selection using copy¶

Field values can be also selected using a filter. A filter is always associated to filter fields, which are the fields where the selection takes place. A feature (record) will be copied only if the filter function returns True. The parameter filter is followed by the layer number; the correspondly field parameter is ffields (filter fields) followed by the layer number. It contais a field name or a list of field names.

def filter1(value):
    return value == 'Bayern'

lrs1 = lrs0.copy('', ofields0=['NAME_4', 'TYPE_4', 'NAME_1'], filter0=filter1, ffields0='NAME_1')

Note also the order of the output fields:

    Bayern
         geom                    NAME_4                 TYPE_4  NAME_1
0     Polygon                    Affing               Gemeinde  Bayern
1     Polygon                   Aichach                  Stadt  Bayern
2     Polygon                  Aindling               Gemeinde  Bayern
...       ...                       ...                    ...     ...
2235  Polygon  Weißenstadter Forst-Nord  Gemeindefreies Gebiet  Bayern
2236  Polygon   Weißenstadter Forst-Süd  Gemeindefreies Gebiet  Bayern
2237  Polygon                 Wunsiedel                  Stadt  Bayern

Using lambda and two filter fields (‘NAME_1’, ‘NAME_4’):

lrs1 = lrs0.copy('', ofields0=['NAME_4', 'TYPE_4'], ffields0=['NAME_1', 'NAME_4'],
                 filter=lambda s, t: s == 'Nordrhein-Westfalen' and t.startswith('A'))
lrs1.show(max_rows=6)

           geom      NAME_4    TYPE_4
 Polygon       Ahaus     Stadt
 Polygon   Ascheberg  Gemeinde
 Polygon  Aldenhoven  Gemeinde
..      ...         ...       ...
Polygon  Altenberge  Gemeinde
Polygon       Ahlen     Stadt
Polygon       Alpen  Gemeinde

Attribute selection using filters¶

Attribute analysis refeers to the selection of layer field values, excluding geometry fields.

print lrs.get_feature_count()  # returns 11302
lrs.set_attribute_filter("NAME_1 = 'Nordrhein-Westfalen'")
print lrs.get_feature_count()  # returns 396
lrs.set_attribute_filter("NAME_1 = 'Bayern'")
print lrs.get_feature_count()  # returns 2238
lrs.set_attribute_filter("NAME_1 = 'Nordrhein-Westfalen' OR NAME_1 = 'Bayern'")
print lrs.get_feature_count()  # returns 2634

Spatial search¶

Spatial search selects a subset of features (records) from one layer in Layers according to specified geometric relationships (inside, intersecting, touching, etc.) to features from another layer.

The method spatial_filter is a generator returning a feature filtered by the spatial operator. The default operator is ‘Intersects’. The feature object should be used only inside the loop.

lrs0 = LayersReader('D:/tmp/DEU_adm_shp/DEU_adm4.shp')
lrs1 = LayersReader('D:/tmp/DEU_wat/DEU_water_areas_dcw.shp')

lrs0 = LayersReader('D:/tmp/DEU_adm_shp/DEU_adm4.shp')
lrs1 = LayersReader('D:/tmp/DEU_wat/DEU_water_areas_dcw.shp')

cities_intersecting_waters = [f.GetField('NAME_4') for f in lrs0.spatial_filter(lrs1)]
print 'Number of cities intersecting waters: {}'.format(len(cities_intersecting_waters))

cities_containing_waters = [f.GetField('NAME_4') for
                            f in lrs0.spatial_filter(lrs1, method='Contains')]
print 'Number of cities containing waters: {}'.format(len(cities_containing_waters))

The method spatial_filter_to_layer returns a filtered copy of the original layer. If no output filename is given, the returned LayersWriter object is stored in RAM. The example below demonstrates the use of the available operators (Intersects, Equals, Disjoint, Touches, Crosses, Within, Contains, and Overlaps). It checks 11302 cities against 907 water bodies.

lrs0 = LayersReader('D:/tmp/DEU_adm_shp/DEU_adm4.shp')
lrs1 = LayersReader('D:/tmp/DEU_wat/DEU_water_areas_dcw.shp')
print lrs0.get_feature_count()
methods = ['Intersects', 'Equals', 'Disjoint', 'Touches', 'Crosses', 'Within',
           'Contains', 'Overlaps']
for method in methods:
    beg = time.time()
    fids = [feat.GetFID() for feat in lrs0.spatial_filter(lrs1, method=method)]
    print '{:<11}{}  {:0.2f} seconds'.format(
        method, str(len(fids)).rjust(5, ' '), time.time() - beg)

print lrs0.get_feature_count()
for method in methods:
    beg = time.time()
    lrs_out = lrs0.spatial_filter_to_layer(lrs1, method=method)
    print '{:<11}{}  {:0.2f} seconds'.format(
        method, str(lrs_out.get_feature_count()).rjust(5, ' '), time.time() - beg)

Outputs:

11302
Intersects  1195  4.12 seconds
Equals         0  2.04 seconds
Disjoint    3230  3.88 seconds
Touches        0  4.35 seconds
Crosses        0  4.40 seconds
Within         0  3.82 seconds
Contains     307  2.68 seconds
Overlaps    1028  4.09 seconds
11302
Intersects  1195  4.24 seconds
Equals         0  2.20 seconds
Disjoint    3230  4.06 seconds
Touches        0  4.76 seconds
Crosses        0  4.96 seconds
Within         0  4.37 seconds
Contains     307  3.24 seconds
Overlaps    1028  4.76 seconds