HIVE_CANNOT_OPEN_SPLIT : 不支持第 <column_name> 列类型 null

Question

HIVE_CANNOT_OPEN_SPLIT：打开 Hive 拆分 s3 时出错：//path/to/file/<>。snappy.parquet：不支持列 ai.ja 类型 null

这只会在我定义一个 "JA" 列时发生，它是一个字符串结构。如果我将该列排除在外，我可以毫无问题地进行查询。架构信息是使用 Apache Spark 从我们的镶木地板文件中获取的。

 The create table statement I'm using to reproduce the error follows:
  CREATE EXTERNAL TABLE <<tablename>>(`ai` struct < acs : varchar(100), ltc : varchar(100), primaryapplicant : struct < bwh : varchar(10), citizenship : varchar(20), currentaddresscity : varchar(50), currentaddressstate : varchar(50), currentaddressstreet2 : varchar(50), ss : varchar(50)>, JA : array < struct < dateofbirth : varchar(50), emailaddress : varchar(50), firstname : varchar(50), lastname : varchar(50), ss : varchar(50)>>, status : varchar(50), uri : varchar(50)>, `pr` struct < pc : struct < cn : varchar(50)>>, `product` array < struct < at : varchar(20), pi : varchar(50), pmn : varchar(256)>>, `ipt` varchar(40) ) PARTITIONED BY ( `owner` varchar(40) ) ROW FORMAT SERDE 'org.apache.hadoop.hive.ql.io.parquet.serde.ParquetHiveSerDe' STORED AS INPUTFORMAT 'org.apache.hadoop.hive.ql.io.parquet.MapredParquetInputFormat' OUTPUTFORMAT 'org.apache.hadoop.hive.ql.io.parquet.MapredParquetOutputFormat' LOCATION 's3://<location>' TBLPROPERTIES ( 'compression_type' = 'snappy', 'numRows' = '2', 'transient_lastDdlTime' = <> )

正在从 parquet 文件中读取。

 Parquet schema : 
  root
 |-- ai: struct (nullable = true)
 |    |-- acs: string (nullable = true)
 |    |-- JA: struct (nullable = true)
 |    |    |-- DateOfBirth: string (nullable = true)
 |    |    |-- EmailAddress: string (nullable = true)
 |    |    |-- FirstName: string (nullable = true)
 |    |    |-- LastName: string (nullable = true)
 |    |    |-- ss: string (nullable = true)
 |    |-- ltc: string (nullable = true)
 |    |-- PrimaryApplicant: struct (nullable = true)
 |    |    |-- bwh: string (nullable = true)
 |    |    |-- Citizenship: string (nullable = true)
 |    |    |-- CurrentAddressCity: string (nullable = true)
 |    |    |-- CurrentAddressState: string (nullable = true)
 |    |    |-- CurrentAddressStreet2: string (nullable = true)
 |    |    |-- ss: string (nullable = true)
 |    |-- Status: string (nullable = true)
 |    |-- uri: string (nullable = true)
 |-- pr: struct (nullable = true)
 |    |-- pc: struct (nullable = true)
 |    |    |-- cn: string (nullable = true)
 |-- Product: array (nullable = true)
 |    |-- element: struct (containsNull = true)
 |    |    |-- at: string (nullable = true)
 |    |    |-- pi: string (nullable = true)
 |    |    |-- pmn: string (nullable = true)
 |-- ipt: string (nullable = true)

此 link https://forums.aws.amazon.com/thread.jspa?threadID=246551 上提出了同样的问题。但仍然无法弄清楚。

有人能帮忙吗？

Answer 1

此问题已解决。

要创建 Athena table，每个字段都应准确映射到架构，即每个字段的顺序应与架构的顺序相同。

HIVE_CANNOT_OPEN_SPLIT : 不支持第 <column_name> 列类型 null

HIVE_CANNOT_OPEN_SPLIT : Column <column_name> type null not supported

apache-spark

parquet

presto

amazon-athena