[FLINK-18678] Hive connector fails to create vector orc reader if user specifies incorrect hive version

• Type: Improvement
• Status: Closed
• Priority: Major
• Resolution: Fixed
• Affects Version/s: 1.11.1
• Fix Version/s: 1.11.2
• Component/s: Connectors / Hive, Formats (JSON, Avro, Parquet, ORC, SequenceFile)
• Labels: pull-request-available
• Assignee: Rui Li
• Reporter: Rui Li
• Created: 23/Jul/20 03:51
• Updated: 27/Apr/21 18:06
• Resolved: 10/Aug/20 03:44

Description

Issue reported by user. User's Hive deployment is 2.1.1 and uses flink-sql-connector-hive-2.2.0_2.11-1.11.0.jar in Flink lib. If user specifies Hive version as 2.1.1, then creating vectorized orc reader fails with exception:

java.lang.ClassCastException: org.apache.hadoop.hive.ql.io.orc.ReaderImpl cannot be cast to org.apache.orc.Reader
  at org.apache.flink.orc.shim.OrcShimV200.createReader(OrcShimV200.java:63) ~[flink-sql-connector-hive-2.2.0_2.11-1.11.0.jar:1.11.0]
  at org.apache.flink.orc.shim.OrcShimV200.createRecordReader(OrcShimV200.java:98) ~[flink-sql-connector-hive-2.2.0_2.11-1.11.0.jar:1.11.0]
  at org.apache.flink.orc.OrcSplitReader.<init>(OrcSplitReader.java:73) ~[flink-sql-connector-hive-2.2.0_2.11-1.11.0.jar:1.11.0]
  at org.apache.flink.orc.OrcColumnarRowSplitReader.<init>(OrcColumnarRowSplitReader.java:54) ~[flink-sql-connector-hive-2.2.0_2.11-1.11.0.jar:1.11.0]
  at org.apache.flink.orc.OrcSplitReaderUtil.genPartColumnarRowReader(OrcSplitReaderUtil.java:91) ~[flink-sql-connector-hive-2.2.0_2.11-1.11.0.jar:1.11.0]
......

Issue Links

• links to: GitHub Pull Request #12988