[IMPALA-2048] Class not found: com.cloudera.impala.hive.serde.ParquetInputFormat error

• Type: Bug
• Status: Closed
• Priority: Blocker
• Resolution: Fixed
• Affects Version/s: Impala 2.2
• Fix Version/s: Impala 2.2.5, Impala 2.3.0
• Labels: catalog-server hive
• Target Version: Impala 2.2.5, Impala 2.3.0
• Assignee: Dimitris Tsirogiannis
• Reporter: Dimitris Tsirogiannis
• Created: 03/Jun/15 21:34
• Updated: 25/Nov/15 18:01
• Resolved: 10/Jun/15 21:17

Description

When Impala is used to compute stats on a partitioned parquet table which was created in hive, the table subsequently becomes unaccessible in Hive and the following error is thrown:

Error: Error while compiling statement: FAILED: SemanticException Class not found: com.cloudera.impala.hive.serde.ParquetInputFormat (state=42000,code=40000)

Steps to reproduce:

1. in beeline: 
>CREATE TABLE sample_partitioned (description STRING, total_emp INT, salary INT) PARTITIONED BY (code string) STORED AS PARQUET; 
>INSERT OVERWRITE TABLE sample_partitioned PARTITION (code) SELECT idh.description, idh.total_emp, idh.salary, idh.code FROM sample_07 idh; 

[sample_partitioned is totally usable now] 

2. in Impala: 
> invalidate metadata; compute stats sample_partitioned; 

[sample_partitioned is still totally usable from impala] 

3. in beeline: 
> select * from sample_partitioned; 
Error: Error while compiling statement: FAILED: SemanticException Class not found: com.cloudera.impala.hive.serde.ParquetInputFormat (state=42000,code=40000)