Regex SerDe 不支持 serialize() 方法错误

Question

我的表结构如下。

CREATE TABLE db.TEST(
f1 string,
f2 string,
f3 string)
ROW FORMAT SERDE
  'org.apache.hadoop.hive.serde2.RegexSerDe'
WITH SERDEPROPERTIES (
  'input.regex'='(.{2})(.{3})(.{4})' )
STORED AS INPUTFORMAT
  'org.apache.hadoop.mapred.TextInputFormat'
OUTPUTFORMAT
  'org.apache.hadoop.hive.ql.io.HiveIgnoreKeyTextOutputFormat'
LOCATION
  'hdfs://nameservice1/location/TEST';

我尝试在表中插入一条记录，如下所示。

insert overwrite table db.TEST2 
select '12' as a , '123' as b , '1234' as c ;

尝试向表中插入数据时，遇到以下错误。

Caused by: java.lang.UnsupportedOperationException: Regex SerDe doesn't support the serialize() method at org.apache.hadoop.hive.serde2.RegexSerDe.serialize(RegexSerDe.java:289)

知道出了什么问题吗？

Answer 1

您使用了错误的 SerDe 类。 org.apache.hadoop.hive.serde2.RegexSerDe不支持序列化。看source code - serialize 方法只抛出 UnsupportedOperationException 异常:

 public Writable serialize(Object obj, ObjectInspector objInspector)
      throws SerDeException {
        throw new UnsupportedOperationException(
          "Regex SerDe doesn't support the serialize() method");
}

解决方案是

使用另一个 SerDe 类: org.apache.hadoop.hive.contrib.serde2.RegexSerDe , 它可以使用 format 序列化行对象字符串。序列化格式应在 SERDEPROPERTIES 中指定。看source code了解更多详情。

SerDe 属性示例:

WITH SERDEPROPERTIES ( 'input.regex' = '(.{2})(.{3})(.{4})','output.format.string' = '%1$2s%2$3s%3$4s') 

对于你的表，它将是这样的:

CREATE TABLE db.TEST(
f1 string,
f2 string,
f3 string)
ROW FORMAT SERDE
  'org.apache.hadoop.hive.contrib.serde2.RegexSerDe'
WITH SERDEPROPERTIES (
  'input.regex'='(.{2})(.{3})(.{4})',
  'output.format.string' = '%1$2s%2$3s%3$4s' )
LOCATION
  'hdfs://nameservice1/location/TEST';