Details
-
Bug
-
Status: Closed
-
Critical
-
Resolution: Fixed
-
3.1.0, 3.1.1, 3.1.2
Description
Scenario is Hive table having same schema as table in Oracle, however when we query the table with data it fails with NPE, below is the trace.
Caused by: java.io.IOException: java.lang.NullPointerException
at org.apache.hadoop.hive.ql.exec.FetchOperator.getNextRow(FetchOperator.java:617) ~[hive-exec-3.1.0.3.1.5.0-152.jar:3.1.0.3.1.5.0-152]
at org.apache.hadoop.hive.ql.exec.FetchOperator.pushRow(FetchOperator.java:524) ~[hive-exec-3.1.0.3.1.5.0-152.jar:3.1.0.3.1.5.0-152]
at org.apache.hadoop.hive.ql.exec.FetchTask.fetch(FetchTask.java:146) ~[hive-exec-3.1.0.3.1.5.0-152.jar:3.1.0.3.1.5.0-152]
at org.apache.hadoop.hive.ql.Driver.getResults(Driver.java:2739) ~[hive-exec-3.1.0.3.1.5.0-152.jar:3.1.0.3.1.5.0-152]
at org.apache.hadoop.hive.ql.reexec.ReExecDriver.getResults(ReExecDriver.java:229) ~[hive-exec-3.1.0.3.1.5.0-152.jar:3.1.0.3.1.5.0-152]
at org.apache.hive.service.cli.operation.SQLOperation.getNextRowSet(SQLOperation.java:473) ~[hive-service-3.1.0.3.1.5.0-152.jar:3.1.0.3.1.5.0-152]
... 34 more
Caused by: java.lang.NullPointerException
at org.apache.hive.storage.jdbc.JdbcSerDe.deserialize(JdbcSerDe.java:164) ~[hive-jdbc-handler-3.1.0.3.1.5.0-152.jar:3.1.0.3.1.5.0-152]
at org.apache.hadoop.hive.ql.exec.FetchOperator.getNextRow(FetchOperator.java:598) ~[hive-exec-3.1.0.3.1.5.0-152.jar:3.1.0.3.1.5.0-152]
at org.apache.hadoop.hive.ql.exec.FetchOperator.pushRow(FetchOperator.java:524) ~[hive-exec-3.1.0.3.1.5.0-152.jar:3.1.0.3.1.5.0-152]
at org.apache.hadoop.hive.ql.exec.FetchTask.fetch(FetchTask.java:146) ~[hive-exec-3.1.0.3.1.5.0-152.jar:3.1.0.3.1.5.0-152]
at org.apache.hadoop.hive.ql.Driver.getResults(Driver.java:2739) ~[hive-exec-3.1.0.3.1.5.0-152.jar:3.1.0.3.1.5.0-152]
at org.apache.hadoop.hive.ql.reexec.ReExecDriver.getResults(ReExecDriver.java:229) ~[hive-exec-3.1.0.3.1.5.0-152.jar:3.1.0.3.1.5.0-152]
at org.apache.hive.service.cli.operation.SQLOperation.getNextRowSet(SQLOperation.java:473) ~[hive-service-3.1.0.3.1.5.0-152.jar:3.1.0.3.1.5.0-152]
... 34 more
Problem appears when column names in Oracle are in Upper case and since in Hive, table and column names are forced to store in lowercase during creation. User runs into NPE error while fetching data.
While deserializing data, input consists of column names in lower case which fails to get the value
rowVal = ((ObjectWritable)value).get();
Log Snio:
=============
2020-07-17T16:49:09,598 INFO [04ed42ec-91d2-4662-aee7-37e840a06036 HiveServer2-Handler-Pool: Thread-104]: dao.GenericJdbcDatabaseAccessor (:()) - Query to execute is [select * from TESTHIVEJDBCSTORAGE] 2020-07-17T16:49:10,642 INFO [04ed42ec-91d2-4662-aee7-37e840a06036 HiveServer2-Handler-Pool: Thread-104]: jdbc.JdbcSerDe (:()) - *** ColumnKey = ID 2020-07-17T16:49:10,642 INFO [04ed42ec-91d2-4662-aee7-37e840a06036 HiveServer2-Handler-Pool: Thread-104]: jdbc.JdbcSerDe (:()) - *** Blob value = {fname=OW[class=class java.lang.String,value=Name1], id=OW[class=class java.lang.Integer,value=1]}
Simple Reproducer for this case.
=============
1. Create table in Oracle
create table TESTHIVEJDBCSTORAGE(ID INT, FNAME VARCHAR(20));
2. Insert dummy data.
Insert into TESTHIVEJDBCSTORAGE values (1, 'Name1');
3. Create JDBCStorageHandler table in Hive.
CREATE EXTERNAL TABLE default.TESTHIVEJDBCSTORAGE_HIVE_TBL (ID INT, FNAME VARCHAR(20)) STORED BY 'org.apache.hive.storage.jdbc.JdbcStorageHandler' TBLPROPERTIES ( "hive.sql.database.type" = "ORACLE", "hive.sql.jdbc.driver" = "oracle.jdbc.OracleDriver", "hive.sql.jdbc.url" = "jdbc:oracle:thin:@orachehostname/XE", "hive.sql.dbcp.username" = "chiran", "hive.sql.dbcp.password" = "supersecurepassword", "hive.sql.table" = "TESTHIVEJDBCSTORAGE", "hive.sql.dbcp.maxActive" = "1" );
4. Query Hive table, fails with NPE.
> select * from default.TESTHIVEJDBCSTORAGE_HIVE_TBL; INFO : Compiling command(queryId=hive_20200717164857_cd6f5020-4a69-4a2d-9e63-9db99d0121bc): select * from default.TESTHIVEJDBCSTORAGE_HIVE_TBL INFO : Semantic Analysis Completed (retrial = false) INFO : Returning Hive schema: Schema(fieldSchemas:[FieldSchema(name:testhivejdbcstorage_hive_tbl.id, type:int, comment:null), FieldSchema(name:testhivejdbcstorage_hive_tbl.fname, type:varchar(20), comment:null)], properties:null) INFO : Completed compiling command(queryId=hive_20200717164857_cd6f5020-4a69-4a2d-9e63-9db99d0121bc); Time taken: 9.914 seconds INFO : Executing command(queryId=hive_20200717164857_cd6f5020-4a69-4a2d-9e63-9db99d0121bc): select * from default.TESTHIVEJDBCSTORAGE_HIVE_TBL INFO : Completed executing command(queryId=hive_20200717164857_cd6f5020-4a69-4a2d-9e63-9db99d0121bc); Time taken: 0.019 seconds INFO : OK Error: java.io.IOException: java.lang.NullPointerException (state=,code=0)
Assuming that there are no repercussions, can we convert the column names to lowercase fetched from Database/Query pointing to table in JDBCStorageHandler?
Attaching the patch for the case.
