DataFrame partitionBy嵌套列

时间:2016-07-12 03:24:47

标签: apache-spark apache-spark-sql spark-dataframe

我试图在嵌套字段上调用partitionBy,如下所示:

val rawJson = sqlContext.read.json(filename)
rawJson.write.partitionBy("data.dataDetails.name").parquet(filenameParquet)

运行时出现以下错误。我确实看到“名称”列为以下架构中的字段。是否有不同的格式来指定嵌套的列名?

  

java.lang.RuntimeException:在模式StructType中找不到分区列data.dataDetails.name(StructField(name,StringType,true),StructField(time,StringType,true),StructField(data,StructType(StructField)(dataDetails, StructType(StructField(name,StringType,true),StructField(id,StringType,true),true)),true))

这是我的json文件:

{  
  "name": "AssetName",
  "time": "2016-06-20T11:57:19.4941368-04:00",
  "data": {
    "type": "EventData",
    "dataDetails": {
      "name": "EventName"
      "id": "1234"

    }
  }
} 

2 个答案:

答案 0 :(得分:2)

这似乎是此处列出的已知问题:https://issues.apache.org/jira/browse/SPARK-18084

我也遇到了这个问题,为了解决这个问题,我能够在我的数据集上取消嵌套列。我的数据集与您的数据集略有不同,但这是策略......

原Json:

{  
  "name": "AssetName",
  "time": "2016-06-20T11:57:19.4941368-04:00",
  "data": {
    "type": "EventData",
    "dataDetails": {
      "name": "EventName"
      "id": "1234"

    }
  }
} 

修改了Json:

{  
  "name": "AssetName",
  "time": "2016-06-20T11:57:19.4941368-04:00",
  "data_type": "EventData",
  "data_dataDetails_name" : "EventName",
  "data_dataDetails_id": "1234"
  }
} 

获取修改Json的代码:

def main(args: Array[String]) {
  ...

  val data = df.select(children("data", df) ++ $"name" ++ $"time"): _*)

  data.printSchema

  data.write.partitionBy("data_dataDetails_name").format("csv").save(...)
}

def children(colname: String, df: DataFrame) = {
  val parent = df.schema.fields.filter(_.name == colname).head
  val fields = parent.dataType match {
    case x: StructType => x.fields
    case _ => Array.empty[StructField]
  }
  fields.map(x => col(s"$colname.${x.name}").alias(s"$colname" + s"_" + s"${x.name}"))
}

答案 1 :(得分:0)

由于此功能自Spark 2.3.1起不可用,因此有一种解决方法。确保处理嵌套字段和根级别的字段之间的名称冲突。

{"date":"20180808","value":{"group":"xxx","team":"yyy"}}
df.select("date","value.group","value.team")
      .write
      .partitionBy("date","group","team")
      .parquet(filenameParquet)

分区最终像

date=20180808/group=xxx/team=yyy/part-xxx.parquet