Marek Pasieka created HIVE-13386:
------------------------------------
Summary: NPE when trying to read parquet file with null values
created by Spark
Key: HIVE-13386
URL: https://issues.apache.org/jira/browse/HIVE-13386
Project: Hive
Issue Type: Bug
Reporter: Marek Pasieka
Trying to "select * from table limit 1" from hive table throws:
Failed with exception java.io.IOException:java.lang.NullPointerException
Steps to reproduce:
1. Run spark sql/streaming job, write dataframe with fields that CONTAIN null
values (null instead of String/null instead of number) as parquet file.
2. Create external partitioned hive table on top of the parquet files. MSCK
repair table.
3. select query returns NPE as described.
Not sure if partitioning is relevant, but the tabled is partitioned by string
value.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
