David Kay created AVRO-1850:
-------------------------------
Summary: Align JSON and binary record serialization
Key: AVRO-1850
URL: https://issues.apache.org/jira/browse/AVRO-1850
Project: Avro
Issue Type: Improvement
Reporter: David Kay
The documentation describes the encoding of Avro records as:
bq.Binary: A record is encoded by encoding the values of its fields in the
order that they are declared. In other words, a record is encoded as just the
concatenation of the encodings of its fields. Field values are encoded per
their schema.
bq.JSON: Except for unions, the JSON encoding is the same as is used to encode
field default values.
The _field default values_ table says that records and maps are both encoded as
JSON type _object_.
*Enhancement:*
There is currently no way to write an Avro schema describing a JSON array of
positional parameters (i.e. an array containing variables of possibly different
type). An Avro record is the datatype representing an ordered collection of
values. For consistency with the binary encoding, and to allow Avro to
represent a schema for JSON tuples, encoding should say:
bq.JSON: Except for unions and records, the JSON encoding is the same as is
used to encode field default values. A record is encoded as an array by
encoding the values of its fields in the order that they are declared.
For the example schema:
{noformat}
{"namespace": "example.avro",
"type": "record",
"name": "User",
"fields": [
{"name": "name", "type": "string"},
{"name": "favorite_number", "type": ["int", "null"]},
{"name": "favorite_color", "type": ["string", "null"]}
]
}
{noformat}
the JSON encoding currently converts an Avro record to an Avro map (JSON
object):
{noformat}
{ "name": "Joe",
"favorite_number": 42,
"favorite_color": null }
{noformat}
Instead Avro records should be encoded in JSON in the same manner as they are
encoded in binary, as a JSON array containing the fields in the order they are
defined:
{noformat}
["Joe", 42, null]
{noformat}
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
