Going from Python data structure to binary avro representation and from avro to dict.

avro_to_dict.py

from io import BytesIO
from avro.io import DatumReader
from avro.datafile import DataFileReader

blob = ... 
schema = ... 

byte_stream = BytesIO(blob)
reader = DataFileReader(byte_stream, DatumReader(schema))
value = [item for item in reader][0]

# value is now a Python dictionary representation of the avro object.

data_to_avro.py

from io import BytesIO
from avro.io import DatumWriter
from avro.datafile import DataFileWriter


data = ... # data is whatever you want to encode
schema = ... 

byte_stream = BytesIO()
writer = DatumWriter()
file_writer = DataFileWriter(byte_stream, writer, schema)
file_writer.append(data)
file_writer.flush()
binary_data = byte_stream.getvalue()
file_writer.close()

# binary_data is now an avro object which you can ship to e.g. Kafka

This is really useful! Thanks!

This is the part I was missing file_writer.flush(): without it when you do byte_stream.getvalue() byte stream is empty.