[HIVE-6056] The AvroSerDe gives out BadSchemaException if a partition is added to the table

• Type: Bug
• Status: Patch Available
• Priority: Major
• Resolution: Unresolved
• Affects Version/s: 0.11.0
• Component/s: Database/Schema

Environment

amazon EMR (hadoop Amazon 1.0.3), avro-1.7.5

• Tags: avro, hive
• Assignee: Ashish Singh
• Reporter: Rushil Gupta
• Created: 18/Dec/13 20:31
• Updated: 08/Oct/14 17:53

Description

While creating an external table if I do not add a partition, I am able to read files using following format:

CREATE external TABLE event
ROW FORMAT SERDE 'org.apache.hadoop.hive.serde2.avro.AvroSerDe'
STORED AS INPUTFORMAT 'org.apache.hadoop.hive.ql.io.avro.AvroContainerInputFormat'
OUTPUTFORMAT 'org.apache.hadoop.hive.ql.io.avro.AvroContainerOutputFormat'
LOCATION 's3n://test-event/input/2013/14/10'
TBLPROPERTIES ('avro.schema.literal' = '..some schema..');

but if I add a partition based on date

CREATE external TABLE event
PARTITIONED BY (ds STRING)
ROW FORMAT SERDE 'org.apache.hadoop.hive.serde2.avro.AvroSerDe'
STORED AS INPUTFORMAT 'org.apache.hadoop.hive.ql.io.avro.AvroContainerInputFormat'
OUTPUTFORMAT 'org.apache.hadoop.hive.ql.io.avro.AvroContainerOutputFormat'
LOCATION 's3n://test-event/input/'
TBLPROPERTIES ('avro.schema.literal' = '..some schema..');

ALTER TABLE event ADD IF NOT EXISTS PARTITION (ds = '2013_12_16') LOCATION '2013/12/16/';

I get the following exception:

java.io.IOException:org.apache.hadoop.hive.serde2.avro.BadSchemaException

Issue Links

• is related to: HIVE-8129 Umbrella JIRA to track Avro improvements
• relates to: HIVE-3953 Reading of partitioned Avro data fails because of missing properties