Hive自建表查询报错：LongColumnVector无法转为DecimalColumnVector

问题 #

使用基于Hadoop 3.3.4、Tez 0.9.2的Hive 3.1.3版本，执行SELECT语句查询Hive自建并管理的表时，查询始终无法完成且报错。核心错误如下：

Caused by: java.lang.ClassCastException: org.apache.hadoop.hive.ql.exec.vector.LongColumnVector cannot be cast to org.apache.hadoop.hive.ql.exec.vector.DecimalColumnVector

从错误看是long与decimal的类型转换问题，但该表由Hive在前序步骤创建并加载/转换数据，若decimal列存在无效值，Hive理应提前报错。且在AWS EMR上使用完全相同的代码和数据未出现此错误，排除数据无效的可能，目前无法定位问题。

表定义如下：

claimid             varchar(50)
claimlineid         int
dos                 date
dosto               date
member              varchar(50)
provider            varchar(50)
setname             varchar(255)
code                varchar(50)
system              varchar(255)
primary             int
positivenegative    int
result              decimal(10,2)
supply              int
size                decimal(10,2)
quantity            decimal(10,2)

完整错误信息：

Vertex failed, vertexName=Map 1, vertexId=vertex_1667735849290_0030_32_15, diagnostics=[Task failed, taskId=task_1667735849290_0030_32_15_000009, diagnostics=[TaskAttempt 0 failed, info=[Error: Error while running task ( failure ) : attempt_1667735849290_0030_32_15_000009_0:java.lang.RuntimeException: java.lang.RuntimeException: Hive Runtime Error while closing operators
        at org.apache.hadoop.hive.ql.exec.tez.TezProcessor.initializeAndRunProcessor(TezProcessor.java:296)
        at org.apache.hadoop.hive.ql.exec.tez.TezProcessor.run(TezProcessor.java:250)
        at org.apache.tez.runtime.LogicalIOProcessorRuntimeTask.run(LogicalIOProcessorRuntimeTask.java:374)
        at org.apache.tez.runtime.task.TaskRunner2Callable$1.run(TaskRunner2Callable.java:73)
        at org.apache.tez.runtime.task.TaskRunner2Callable$1.run(TaskRunner2Callable.java:61)
        at java.security.AccessController.doPrivileged(Native Method)
        at javax.security.auth.Subject.doAs(Subject.java:422)
        at org.apache.hadoop.security.UserGroupInformation.doAs(UserGroupInformation.java:1657)
        at org.apache.tez.runtime.task.TaskRunner2Callable.callInternal(TaskRunner2Callable.java:61)
        at org.apache.tez.runtime.task.TaskRunner2Callable.callInternal(TaskRunner2Callable.java:37)
        at org.apache.tez.common.CallableWithNdc.call(CallableWithNdc.java:36)
        at com.google.common.util.concurrent.TrustedListenableFutureTask$TrustedFutureInterruptibleTask.runInterruptibly(TrustedListenableFutureTask.java:108)
        at com.google.common.util.concurrent.InterruptibleTask.run(InterruptibleTask.java:41)
        at com.google.common.util.concurrent.TrustedListenableFutureTask.run(TrustedListenableFutureTask.java:77)
        at java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1149)
        at java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:624)
        at java.lang.Thread.run(Thread.java:750)
Caused by: java.lang.RuntimeException: Hive Runtime Error while closing operators
        at org.apache.hadoop.hive.ql.exec.tez.MapRecordProcessor.close(MapRecordProcessor.java:488)
        at org.apache.hadoop.hive.ql.exec.tez.TezProcessor.initializeAndRunProcessor(TezProcessor.java:284)
        ... 16 more
Caused by: org.apache.hadoop.hive.ql.metadata.HiveException: org.apache.hadoop.hive.ql.metadata.HiveException: org.apache.hadoop.hive.ql.metadata.HiveException:
DeserializeRead detail: Reading byte[] of length 4096 at start offset 4 for length 100 to read 14 fields with types [varchar(50), int, date, date, varchar(50), varchar(50), varchar(255), varchar(50), varchar(255), int, decimal(10,2), int, decimal(10,2), decimal(10,2)].  Read field #14 at field start position 0 current read offset 104
        at org.apache.hadoop.hive.ql.exec.MapJoinOperator.closeOp(MapJoinOperator.java:611)
        at org.apache.hadoop.hive.ql.exec.vector.mapjoin.VectorMapJoinGenerateResultOperator.closeOp(VectorMapJoinGenerateResultOperator.java:681)
        at org.apache.hadoop.hive.ql.exec.Operator.close(Operator.java:733)
        at org.apache.hadoop.hive.ql.exec.Operator.close(Operator.java:757)
        at org.apache.hadoop.hive.ql.exec.tez.MapRecordProcessor.close(MapRecordProcessor.java:477)
        ... 17 more
Caused by: org.apache.hadoop.hive.ql.metadata.HiveException: org.apache.hadoop.hive.ql.metadata.HiveException:
DeserializeRead detail: Reading byte[] of length 4096 at start offset 4 for length 100 to read 14 fields with types [varchar(50), int, date, date, varchar(50), varchar(50), varchar(255), varchar(50), varchar(255), int, decimal(10,2), int, decimal(10,2), decimal(10,2)].  Read field #14 at field start position 0 current read offset 104
        at org.apache.hadoop.hive.ql.exec.vector.mapjoin.VectorMapJoinGenerateResultOperator.reProcessBigTable(VectorMapJoinGenerateResultOperator.java:609)
        at org.apache.hadoop.hive.ql.exec.MapJoinOperator.continueProcess(MapJoinOperator.java:671)
        at org.apache.hadoop.hive.ql.exec.MapJoinOperator.closeOp(MapJoinOperator.java:604)
        ... 21 more
Caused by: org.apache.hadoop.hive.ql.metadata.HiveException:
DeserializeRead detail: Reading byte[] of length 4096 at start offset 4 for length 100 to read 14 fields with types [varchar(50), int, date, date, varchar(50), varchar(50), varchar(255), varchar(50), varchar(255), int, decimal(10,2), int, decimal(10,2), decimal(10,2)].  Read field #14 at field start position 0 current read offset 104
        at org.apache.hadoop.hive.ql.exec.vector.mapjoin.VectorMapJoinGenerateResultOperator.reProcessBigTable(VectorMapJoinGenerateResultOperator.java:589)
        ... 23 more
Caused by: java.lang.ClassCastException: org.apache.hadoop.hive.ql.exec.vector.LongColumnVector cannot be cast to org.apache.hadoop.hive.ql.exec.vector.DecimalColumnVector
        at org.apache.hadoop.hive.ql.exec.vector.VectorDeserializeRow.storePrimitiveRowColumn(VectorDeserializeRow.java:687)
        at org.apache.hadoop.hive.ql.exec.vector.VectorDeserializeRow.storeRowColumn(VectorDeserializeRow.java:934)
        at org.apache.hadoop.hive.ql.exec.vector.VectorDeserializeRow.deserialize(VectorDeserializeRow.java:1360)
        at org.apache.hadoop.hive.ql.exec.vector.mapjoin.VectorMapJoinGenerateResultOperator.reProcessBigTable(VectorMapJoinGenerateResultOperator.java:585)
        ... 23 more

排查方向与解决方案 #

1. 临时禁用向量查询引擎 #

向量执行引擎的类型匹配逻辑可能存在bug，先临时关闭验证：

set hive.vectorized.execution.enabled=false;
set hive.vectorized.execution.decimal.enabled=false;

执行上述配置后重新运行查询，如果能正常完成，说明问题出在向量执行的decimal处理逻辑上。

2. 对比EMR与当前环境的补丁差异 #

AWS EMR的Hive会集成定制补丁修复已知bug，当前Hive 3.1.3可能存在未修复的向量转换问题。检查Hive官方JIRA中HIVE-25000至HIVE-26000区间的bug报告，查找类似LongColumnVector转DecimalColumnVector的ClassCast问题，若找到对应bug，可手动应用补丁或升级到包含修复的Hive版本。

3. 验证表的存储与序列化配置 #

• 执行DESCRIBE FORMATTED <你的表名>;，对比EMR环境中该表的存储格式（ORC/Parquet/Text）、序列化库（serialization.lib）是否完全一致。
• 如果使用ORC存储，检查ORC文件的元数据是否异常：执行orcfiledump <HDFS上的ORC文件路径>，查看decimal列的类型定义是否与表结构匹配。

4. 禁用自动MapJoin #

错误发生在MapJoin的反序列化阶段，临时关闭自动MapJoin验证：

set hive.auto.convert.join=false;

重新运行查询，若问题消失，说明是MapJoin与向量执行结合时的bug，可调整hive.mapjoin.smalltable.filesize参数，避免触发MapJoin，或等待对应补丁修复。

5. 重新生成表数据 #

即使数据本身有效，前序写入过程中可能出现decimal列的向量存储异常。尝试重新创建表并重新加载数据：

DROP TABLE IF EXISTS <你的表名>;
-- 重新执行建表语句
-- 重新执行数据加载/转换语句

完成后再运行查询，看是否解决问题。

内容的提问来源于stack exchange，提问作者Patrick Tucci