Kartoza - Reading and Writing Shapefiles in Python with Fiona

What is Fiona?

Fiona is a FOSS Python library that can be used to read and write GIS formats like shape file and geopackage. We may stumble upon some use cases where we need to import or export GIS data from shapefiles and this is where Fiona comes to the rescue. Its usage is straightforward, making it simple to use.

Installing Fiona

We can install Fiona easily using

pip install fiona

Reading Shapefiles

Fiona has built in function to open shapefiles. First, import the necessary Fiona function.

import fiona
from fiona.model import to_dict

Then, we open the shapefile using fiona.open(). We can wrap it in the Python context manager using with.

with fiona.open('path/to/shapefile.shp', 'r') as shapefile:
  ...

Inside the code block, we can do things like getting the CRS info.

  ...  
  print(shapefile.crs)

Which in my case will show:

EPSG:4326

When we open a shapefile using Fiona, it will be opened as a Fiona collection object. So when we do

  ...
  print(shapefile)

It will print:

<open Collection 'path/to/shapefile.shp:shapefile', mode 'r' at 0x7fd460aa3df0>

We can convert it to a dictionary using to_dict.

  ...
  print(to_dict(shapefile))

And it will be converted to dictionary.

{'geometry': {'coordinates': (60.0, 60.0), 'type': 'Point'}, 'id': '0', 'properties': {'count': 1, 'name': 'Point 1'}, 'type': 'Feature'}

The final code will look like this

import fiona
from fiona.model import to_dict

with fiona.open('export/my_shapefile.shp', 'r') as shapefile:
    print(to_dict(shapefile))
    for record in shapefile:
        # do something with the records
        ...

Writing Shapefiles

When we want to create a shapefile using Fiona, first we need to import Fiona and from_epsg.

import fiona
from fiona.crs import from_epsg

from_epsg is a shortcut to the CRS mappings from EPSG codes. Then, we have to define the shapefile schema. In this example, a simple schema will be used

schema={
	'geometry': 'Point', 
	'properties': {
  'count': 'int',
  'mean': 'float',
  'name': 'str:10', 
	}
}

That schema means these:

1. Geometry type would be Point
2. The feature would have 3 attributes:
3. count, with type int.
4. mean, with type float.
5. name, with type str and maximum length of 10.

You can read the schema detail (like what field types available) in https://fiona.readthedocs.io/en/stable/manual.html#format-drivers-crs-bounds-and-schema.

After defining the schema, open the destination file (could be existing or non-existing file). Here, we open a non-existing file in write mode using the context manager:

with fiona.open('path/to/shapefile.shp', 'w', crs=from_epsg(4326), driver='ESRI Shapefile', schema=schema) as output:

We also specify these things when opening the file:

1. Set the CRS to EPSG:4326 using from_epsg(4326) which will convert the EPSG integer code to the EPSG mapping.
2. Use 'Esri Shapefile' as the driver.
3. Set the schema to the one we defined previously.

Inside the context manager, we can build the geometry and properties. The geometry needs to be in GeoJSON dictionary format, while properties is simply in dictionary format.

	...
    point = {'type': 'Point', 'coordinates': (60, 60)}
	prop = {
  'mean': 1.375,
  'name': "Point 1",
  'count': 1,
	}

Finally, write the geometry and properties into the shapefile.

    ...
    output.write({'geometry':point,'properties': prop})

The full code should look like this

import fiona
from fiona.crs import from_epsg


# Shapefile schema
schema={
	'geometry': 'Point', 
	'properties': {
  'count': 'int',
  'mean': 'float',
  'name': 'str:10', 
	}
}
with fiona.open('path/to/shapefile.shp', 'w', crs=from_epsg(4326), driver='ESRI Shapefile', schema=schema) as output:
	point = {'type': 'Point', 'coordinates': (60, 60)}
	prop = {
  'mean': 1.375,
  'name': "Point 1",
  'count': 1,
	}
	output.write({'geometry':point,'properties': prop})