Skip to main content

Pandas reader for the BUFR format using ecCodes.

Project description

Pandas reader for the BUFR format using ecCodes.

Features with development status Alpha:

  • extracts observations from a BUFR file as a Pandas DataFrame,
  • reads BUFR 3 and 4 files with uncompressed and compressed subsets,
  • supports all modern versions of Python 3.7, 3.6, 3.5 and PyPy3,
  • works on Linux, MacOS and Windows, the ecCodes C-library is the only binary dependency,
  • PyPI package with no install time build (binds via CFFI ABI mode).


  • no special handling of nodata values, yet,
  • no conda-forge package (yet),
  • filters only match exact values.


The easiest way to install pdbufr binary dependencies is via Conda:

$ conda install -c conda-forge eccodes

and pdbufr itself as a Python package from PyPI with:

$ pip install pdbufr

System dependencies

The Python module depends on the ECMWF ecCodes library that must be installed on the system and accessible as a shared library. Some Linux distributions ship a binary version that may be installed with the standard package manager. On Ubuntu 18.04 use the command:

$ sudo apt-get install libeccodes0

On a MacOS with HomeBrew use:

$ brew install eccodes

As an alternative you may install the official source distribution by following the instructions at

You may run a simple selfcheck command to ensure that your system is set up correctly:

$ python -m pdbufr selfcheck
Found: ecCodes v2.13.1.
Your system is ready.


First, you need a well-formed BUFR file, if you don’t have one at hand you can download our sample file:

$ wget

You can explore the file with ecCodes command line tools bufr_ls and bufr_dump to understand the structure and the keys/values you can use to select the observations you are interested in.

The pdbufr.read_bufr function return a pandas.DataDrame with the requested columns. It accepts query filters on the BUFR message header that are very fast and query filters on the observation keys. Filters match on a single value or on one value in a list and the are always in logical and:

>>> import pdbufr
>>> df_all = pdbufr.read_bufr('temp.bufr', columns=('stationNumber', 'latitude', 'longitude'))
>>> df_all.head()
   stationNumber  latitude  longitude
0            907     58.47     -78.08
1            823     53.75     -73.67
2              9    -90.00       0.00
3            486     18.43     -69.88
4            165     21.98    -159.33

>>> df_one = pdbufr.read_bufr(
...     'temp.bufr',
...     columns=('stationNumber', 'latitude', 'longitude'),
...     filters={'stationNumber': 907},
... )
>>> df_one.head()
   stationNumber  latitude  longitude
0            907     58.47     -78.08

>>> df_two = pdbufr.read_bufr(
...     'temp.bufr',
...     columns=('stationNumber', 'latitude', 'longitude', 'data_datetime', 'pressure', 'airTemperature'),
...     filters={'stationNumber': [823, 9]},
... )

>>> df_two.head()
   stationNumber  latitude  longitude  pressure  airTemperature       data_datetime
0            823     53.75     -73.67  100000.0  -1.000000e+100 2008-12-08 12:00:00
1            823     53.75     -73.67   97400.0    2.567000e+02 2008-12-08 12:00:00
2            823     53.75     -73.67   93700.0    2.551000e+02 2008-12-08 12:00:00
3            823     53.75     -73.67   92500.0    2.553000e+02 2008-12-08 12:00:00
4            823     53.75     -73.67   90600.0    2.567000e+02 2008-12-08 12:00:00

>>> df_two.tail()
     stationNumber  latitude  longitude  pressure  airTemperature       data_datetime
190              9     51.77      36.17    2990.0  -1.000000e+100 2008-12-08 12:00:00
191              9     51.77      36.17    2790.0    2.063000e+02 2008-12-08 12:00:00
192              9     51.77      36.17    2170.0  -1.000000e+100 2008-12-08 12:00:00
193              9     51.77      36.17    2000.0    2.031000e+02 2008-12-08 12:00:00
194              9     51.77      36.17    1390.0    1.979000e+02 2008-12-08 12:00:00


The main repository is hosted on GitHub, testing, bug reports and contributions are highly welcomed and appreciated:

Please see the CONTRIBUTING.rst document for the best way to help.

Lead developer:

Main contributors:

See also the list of contributors who participated in this project.


Copyright 2019 European Centre for Medium-Range Weather Forecasts (ECMWF).

Licensed under the Apache License, Version 2.0 (the “License”); you may not use this file except in compliance with the License. You may obtain a copy of the License at: Unless required by applicable law or agreed to in writing, software distributed under the License is distributed on an “AS IS” BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. See the License for the specific language governing permissions and limitations under the License.

Project details

Release history Release notifications

Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Files for pdbufr, version 0.8.0
Filename, size File type Python version Upload date Hashes
Filename, size pdbufr-0.8.0-py2.py3-none-any.whl (11.5 kB) File type Wheel Python version py2.py3 Upload date Hashes View hashes
Filename, size pdbufr-0.8.0.tar.gz (2.7 MB) File type Source Python version None Upload date Hashes View hashes

Supported by

Elastic Elastic Search Pingdom Pingdom Monitoring Google Google BigQuery Sentry Sentry Error logging AWS AWS Cloud computing DataDog DataDog Monitoring Fastly Fastly CDN SignalFx SignalFx Supporter DigiCert DigiCert EV certificate StatusPage StatusPage Status page