Skip to main content

Fast read/write of AVRO files

Project description

# fastavro [![Build Status](https://travis-ci.org/tebeka/fastavro.svg?branch=master)](https://travis-ci.org/tebeka/fastavro) [![Documentation Status](https://readthedocs.org/projects/fastavro/badge/?version=latest)](http://fastavro.readthedocs.io/en/latest/?badge=latest)

Because the Apache Python avro package is written in pure Python, it is relatively slow. In one test case, it takes about 14 seconds to iterate through a file of 10,000 records. By comparison, the JAVA avro SDK reads the same file in 1.9 seconds.

The fastavro library was written to offer performance comparable to the Java library. With regular CPython, fastavro uses C extensions which allow it to iterate the same 10,000 record file in 1.7 seconds. With PyPy, this drops to 1.5 seconds (to be fair, the JAVA benchmark is doing some extra JSON encoding/decoding).

fastavro supports the following Python versions:

  • Python 2.7

  • Python 3.4

  • Python 3.5

  • Python 3.6

  • PyPy

  • PyPy3

[Cython]: http://cython.org/

# Documentation

Documentation is available at http://fastavro.readthedocs.io/en/latest/

# Installing fastavro is available both on [PyPi](http://pypi.python.org/pypi)

pip install fastavro

and on [conda-forge](https://conda-forge.github.io) conda channel.

conda install -c conda-forge fastavro

# Hacking

As recommended by Cython, the C files output is distributed. This has the advantage that the end user does not need to have Cython installed. However it means that every time you change fastavro you need to run make.

### Releasing

We release both to [pypi][pypi] and to [conda-forge][conda-forge].

We assume you have [twine][twine] installed and that you’ve created your own fork of [fastavro-feedstock][feedstock].

  • Make sure the tests pass

  • Copy the windows build artifacts for the new version from https://ci.appveyor.com/project/scottbelden/fastavro to the dist folder

  • Run make publish

  • Note the sha signature emitted at the above

  • Switch to feedstock directory and edit recipe/meta.yaml
    • Update version and sha256 variables at the top of the file

    • Run python recipe/test_recipe.py

    • Submit a [PR][pr]

[conda-forge]: https://conda-forge.org/ [feedstock]: https://github.com/conda-forge/fastavro-feedstock [pr]: https://conda-forge.org/#update_recipe [pypi]: https://pypi.python.org/pypi [twine]: https://pypi.python.org/pypi/twine

# Changes

See the [ChangeLog]

[ChangeLog]: https://github.com/tebeka/fastavro/blob/master/ChangeLog

# Contact

[Project Home](https://github.com/tebeka/fastavro)

Project details


Release history Release notifications | RSS feed

Download files

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

Source Distribution

fastavro-0.17.10.tar.gz (443.0 kB view hashes)

Uploaded Source

Built Distributions

fastavro-0.17.10-cp36-cp36m-win_amd64.whl (278.0 kB view hashes)

Uploaded CPython 3.6m Windows x86-64

fastavro-0.17.10-cp36-cp36m-manylinux1_x86_64.whl (1.2 MB view hashes)

Uploaded CPython 3.6m

fastavro-0.17.10-cp35-cp35m-win_amd64.whl (270.9 kB view hashes)

Uploaded CPython 3.5m Windows x86-64

fastavro-0.17.10-cp35-cp35m-manylinux1_x86_64.whl (1.2 MB view hashes)

Uploaded CPython 3.5m

fastavro-0.17.10-cp34-cp34m-manylinux1_x86_64.whl (1.2 MB view hashes)

Uploaded CPython 3.4m

fastavro-0.17.10-cp27-cp27m-win_amd64.whl (276.1 kB view hashes)

Uploaded CPython 2.7m Windows x86-64

fastavro-0.17.10-cp27-cp27m-manylinux1_x86_64.whl (1.1 MB view hashes)

Uploaded CPython 2.7m

Supported by

AWS AWS Cloud computing and Security Sponsor Datadog Datadog Monitoring Fastly Fastly CDN Google Google Download Analytics Microsoft Microsoft PSF Sponsor Pingdom Pingdom Monitoring Sentry Sentry Error logging StatusPage StatusPage Status page