Home | About | Sematext search-lucene.com search-hadoop.com
 Search Hadoop and all its subprojects:

Switch to Plain View
Avro >> mail # user >> Does Avro Serde support schema evolution?


Copy link to this message
-
Does Avro Serde support schema evolution?
Hi,

Assume v0 schema is:

{"namespace": "example.avro",
 "type": "record",
 "name": "User",
 "fields": [
     {"name": "name", "type": "string"},
     {"name": "favorite_number",  "type": ["int", "null"]},
     {"name": "favorite_color", "type": ["string", "null"]}
 ]
}

And v1 schema is (just added an optional field named age):

{"namespace": "example.avro",
 "type": "record",
 "name": "User",
 "fields": [
     {"name": "name", "type": "string"},
     {"name": "favorite_number",  "type": ["int", "null"]},
     {"name": "favorite_color", "type": ["string", "null"]},
     {"name": "age",  "type": ["int", "null"]}
 ]
}

I created a table using v1 user schema on an Avro file written with v0 of
the user schema. When running a select * on the table I get an error

Failed with exception java.io.IOException:org.apache.avro.AvroTypeException:
Found {...}

Isn't this a scenario that Avro and by extension the AvroSerde should
support?
--
View this message in context: http://apache-avro.679487.n3.nabble.com/Does-Avro-Serde-support-schema-evolution-tp4028398.html
Sent from the Avro - Users mailing list archive at Nabble.com.
+
kulkarni.swarnim@...) 2013-10-17, 20:56
+
kva 2013-10-18, 19:09