在hive中使用parquet (CDH4.3)

hadoop版本 cdh4.3

使用impala创建parquet表后,查询会出错。

[impala:21000] > SELECT * FROM foo;
Query: SELECT * FROM foo
ERROR: AnalysisException: Failed TO LOAD metadata FOR TABLE: DEFAULT.foo
CAUSED BY: TableLoadingException: Failed TO LOAD metadata FOR TABLE: foo
CAUSED BY: MetaException: org.apache.hadoop.hive.serde2.SerDeException SerDe parquet.hive.serde.ParquetHiveSerDe does NOT exist

原因是hive并没有这些lib,下载它们并放入/opt/cloudera/parcels/CDH/lib/hive/lib目录(我是使用cloudera manager部署的),创建脚本下载

#!/bin/sh
#parquet-pig parquet-scrooge parquet-test-hadoop2 parquet-thrift parquet-avro parquet-cascading 
 
for f in parquet-column parquet-common parquet-encoding parquet-generator parquet-hadoop parquet-hive 
do
curl -O http://repo1.maven.org/maven2/com/twitter/${f}/1.2.4/${f}-1.2.4.jar
#curl -O http://oss.sonatype.org/service/local/repositories/releases/content/com/twitter/${f}/1.2.4/${f}-1.2.4.jar
done
curl -O http://repo1.maven.org/maven2/com/twitter/parquet-format/1.0.0/parquet-format-1.0.0.jar

然后把他们拷贝进去

cp parquet-*  /opt/cloudera/parcels/CDH/lib/hive/lib

可能要重启metastore,然后在impala中刷新metastore

INVALIDATE METADATA;

在impala修改parquet表

create table test2 (name STRING) STORED AS PARQUETFILE;

插入数据

insert into test2 select * from test;

参考

https://issues.cloudera.org/browse/IMPALA-574

 




fatkun

没有评论


You can leave the first : )



发表评论

电子邮件地址不会被公开。