CODEWITHSUNDEEP
hadoopavro
James
James

Reputation: 1461

Test and convert avro schema (.avsc) to .avro. AttributeError, array and encoding

I'm just starting hadoop, and I'm using Avro (fastavro).

1- I want to validate schema and convert to .avro file.

{
 "type": "record",
 "name": "Node",
 "fields": [
    {
        "name": "nom",
        "type": "string"
    },
    {
        "name": "zone",
        "type": {
            "type": "map",
            "values": "string"
        }
    },
    {
        "name": "price",
        "type": "float"
    },
    {
        "name": "type",
        "type": "string"
    }
  ]
}

My test file (validate schema) : 

#!/usr/local/bin/python
# -*- coding: utf-8 -*-
import json
import fastavro

schema = json.load(open("myschema.avsc"))
records = [
    {
        "nom": "blabla",
        "zone": ["north", "south", "east"],
        "prix": 4.0,
        "type": "geoloc"
    }
]

fastavro.writer(open("myschema.avro", "wb"), schema, records)

I've this error :

Traceback (most recent call last):
  File "test-schema.py", line 17, in <module>
    fastavro.writer(open("myschema.avro", "wb"), schema, records)
  File "/var/www/data-machine/HDFS/env/lib/python3.5/site-packages/fastavro/writer.py", line 614, in writer
    output.write(record)
  File "/var/www/data-machine/HDFS/env/lib/python3.5/site-packages/fastavro/writer.py", line 537, in write
    write_data(self.io, record, self.schema)
  File "/var/www/data-machine/HDFS/env/lib/python3.5/site-packages/fastavro/writer.py", line 432, in write_data
    return fn(fo, datum, schema)
  File "/var/www/data-machine/HDFS/env/lib/python3.5/site-packages/fastavro/writer.py", line 363, in write_record
    name, field.get('default')), field['type'])
  File "/var/www/data-machine/HDFS/env/lib/python3.5/site-packages/fastavro/writer.py", line 432, in write_data
    return fn(fo, datum, schema)
  File "/var/www/data-machine/HDFS/env/lib/python3.5/site-packages/fastavro/writer.py", line 232, in write_map
    for key, val in iteritems(datum):
  File "/var/www/data-machine/HDFS/env/lib/python3.5/site-packages/fastavro/six.py", line 27, in py3_iteritems
    return obj.items()
AttributeError: 'list' object has no attribute 'items'

2- And, if I add an array : 

{
    "name": "ingredients", 
    "type": ["string"]
},

The error : 

  File "/var/www/data-machine/HDFS/env/lib/python3.5/site-packages/fastavro/writer.py", line 345, in write_union
    raise ValueError(msg)
ValueError: ["north", "south", "east"] (type <class 'list'>) do not match ['string']

And, finishing, I want to make the "zone" field optionnal...

Thanks :) Fabrice

Upvotes: 0

Views: 2381

Answers (1)

hlagos
hlagos

Reputation: 7947

Your record information for the map is wrong. It is expecting something like

"zone":{"key1":"val1","key2":"val2","key3":"val3"},

it is a map, not a set. If you want something like your example, you will need to use an array instead of a map

Upvotes: 1

Related Questions