Skip to main content

1 Minute Introduction to cavro


Install cavro using pip:

$ pip install cavro


$ python3 -m pip install cavro

Basic Usage

Decoding AVRO objects

If you have a schema and an avro object in bytes, then decoding it is simple:

import cavro

schema = cavro.Schema('{"type": "int"}')
encoded_avro = b'\x8a1'

decoded = schema.binary_decode(encoded_avro)
assert decoded == 3141

Encoding AVRO objects

Encoding values to avro is the opposite:

import cavro

schema = cavro.Schema('{"type": "int"}')
value = 3141

encoded = schema.binary_encode(value)
assert encoded == b'\x8a1'

Reading & Writing Files

cavro supports reading and writing avro binary content from files, both as raw avro objects, and from the avro object container format.

Decoding Raw AVRO from a file

If you have a file (or file-like object) containing the avro data, this can be decoded directly.

cavro just reads a single value from the stream, and does not seek, or check if any more objects are on the stream. Serial reading/writing and checking for extra data can easily be added in the calling code.

import cavro

schema = cavro.Schema('{"type": "long"}')

with open('my-file.bin', 'rb') as fh:
reader = cavro.FileReader(fh)

Encoding Raw AVRO to a file

import cavro

schema = cavro.Schema({"type": "long"})

with open('my-file.bin', 'wb') as fh:
writer = cavro.FileWriter(fh)
schema.binary_write(writer, 3141)

Reading AVRO object container files

Files that are in the avro object container format can be read directly:

import cavro

for obj in cavro.ContainerReader('file.avro'):

Writing AVRO object container files

To write a container format file:

import cavro

schema = cavro.Schema(...)
with cavro.ContainerWriter('file.avro', schema, codec='snappy') as writer: