Create new shapefile based on values ranges in existing shapefile using gdal/ogr/python

Question

I have a shapefile containing many thousands of polygons with values stored in an attribute field and ranging 0-105. I would like to create a much simplified file that contains classes for value ranges. E.g. values 30-50 = class 1, values 50-70 = class 2, values 70-100 = class 3. The classes can be integers. Would it then be possible to simplify this to a shapefile containing 3 feature classes (class 1, class 2, class 3) each class containing every polygon fitting those criteria?

There are no overlapping polygons. Attributes from the original shapefile do not need to be preserved. I'd like to use python and ogr to accomplish this.

Answer 1

You can use PyQGIS code (I know that you use Ubuntu) at the Python Console of QGIS. To test my sugestion, I used the next code where a QgsExpression object has your class 1 values:

layer = iface.activeLayer()

expression = QgsExpression( u'"values" >= 30 AND "values" <= 50' )

idx = [ feat.attributes()[0] for feat in layer.getFeatures() 
        if expression.evaluate( feat ) ]

feats = [ feat for feat in layer.getFeatures() 
          if expression.evaluate( feat )]

for feat in feats:
    area = feat.geometry().area()
    print u"area{:.0f} = {:.2f}".format(idx[feats.index(feat)], area)

crs = layer.crs()

epsg = crs.postgisSrid()

uri = "Polygon?crs=epsg:" + str(epsg) + "&field=id:integer""&field=area&index=yes"

new_layer = QgsVectorLayer(uri, 
                           'class1', 
                           'memory')

QgsMapLayerRegistry.instance().addMapLayer(new_layer)

prov = new_layer.dataProvider()

n = len(feats)

for i in range(n):

    feats[i].setAttributes([i, feats[i].geometry().area()])

prov.addFeatures(feats)

I ran this code with the polygon vector layer of next image; where you can observe a 'values' field at the attributes table.

Result was a new memory layer (blue color) with the features that match the query for your class.

Areas were calculated for two different methods for corroborating this approach (features were selected as I expected).