PDAL provides Python support in two significant ways. First it embeds Python to allow you to write Python programs that interact with data using filters.python filter. Second, it extends Python by providing an extension that Python programmers can use to leverage PDAL capabilities in their own applications.


PDAL’s Python story always revolves around Numpy support. PDAL’s data is provided to both the filters ands the extension as Numpy arrays.


PDAL supports both Python 2.7 and Python 3.4+. Continuous Integration tests Python 2.7 on both Linux and Windows. Python 3 is used by a number of developers for adhoc development and testing.


PDAL allows users to embed Python functions inline with other Pipeline processing operations. The purpose of this capability is to allow users to write small programs that implement interesting actions without requiring a full C++ development activity of building a PDAL stage to implement it. A Python filter is an opportunity to interactively and iteratively prototype a data operation without strong considerations of performance or generality. If something works well enough, maybe one takes on the effort to formalize it, but that isn’t necessary. PDAL’s embed of Python allows you to be as grimy as you need to get the job done.


Embedding a Python function to take Z values read from a readers.las and then output them to a writers.bpf.


PDAL provides a Python extension that gives users access to executing Pipeline instantiations and capturing the results as Numpy arrays. This mode of operation is useful if you are looking to have PDAL simply act as your data format and processing handler.

Python extension users are expected to construct their own JSON Pipeline using Python’s json library, or whatever other libraries they wish to manipulate JSON. They then feed it into the extension and get back the results as Numpy arrays:

json = """
  "pipeline": [
        "type": "filters.sort",
        "dimension": "X"

import pdal
pipeline = pdal.Pipeline(json)
pipeline.validate() # check if our JSON and options were good
pipeline.loglevel = 8 #really noisy
count = pipeline.execute()
arrays = pipeline.arrays
metadata = pipeline.metadata
log = pipeline.log


PDAL Python bindings require a working PDAL install (PDAL) and then installation of the Python extension. The extension lives on PyPI at and you should use that version as your canonical Python extension install.

Install from local

In the source code of PDAL there is a python folder, you have to enter there and run

python build
# this should be run as administrator/super user
python install

Install from repository

The second method to install the PDAL Python extension is to use pip or easy_install, you have to run the command as administrator.

pip install PDAL


To install pip please read here

Install from Conda

The final method to install the PDAL Python extension is to use conda. An added advantage of using Conda to install the extension is that Conda will also install PDAL.

conda install -c conda-forge python-pdal


The official pdal and python-pdal packages reside in the conda-forge channel, which can be added via conda config or manually specified with the -c option, as shown in the examples below.

It is recommended that you actually either install PDAL and the Python extension either into an existing environment

conda install -n <environment name> -c conda-forge python-pdal

or create a new environment from scratch

conda create -n <environment name> -c conda-forge python-pdal

Once the environment has been created, you will be prompted to activate it.

conda activate <environment name>