如何使用Apache Spark 1.4.0数据框提取复杂的JSON结构

Question

我正在使用新的Apache Spark版本1.4.0数据帧API从Twitter的状态JSON中提取信息,主要集中在实体对象上 - 此问题的相关部分如下所示:

{
  ...
  ...
  "entities": {
    "hashtags": [],
    "trends": [],
    "urls": [],
    "user_mentions": [
      {
        "screen_name": "linobocchini",
        "name": "Lino Bocchini",
        "id": 187356243,
        "id_str": "187356243",
        "indices": [ 3, 16 ]
      },
      {
        "screen_name": "jeanwyllys_real",
        "name": "Jean Wyllys",
        "id": 111123176,
        "id_str": "111123176",
        "indices": [ 79, 95 ]
      }
    ],
    "symbols": []
  },
  ...
  ...
}

对于如何从基元类型作为提取信息的几个例子string,integer等等-但我无法找到如何处理这些类型的任何复杂的结构.

我尝试了下面的代码,但它仍然无法正常工作,它会引发异常

val sqlContext = new org.apache.spark.sql.hive.HiveContext(sc)

val tweets = sqlContext.read.json("tweets.json")

// this function is just to filter empty entities.user_mentions[] nodes
// some tweets doesn't contains any mentions
import org.apache.spark.sql.functions.udf
val isEmpty = udf((value: List[Any]) => value.isEmpty)

import org.apache.spark.sql._
import sqlContext.implicits._
case class UserMention(id: Long, idStr: String, indices: Array[Long], name: String, screenName: String)

val mentions = tweets.select("entities.user_mentions").
  filter(!isEmpty($"user_mentions")).
  explode($"user_mentions") {
  case Row(arr: Array[Row]) => arr.map { elem =>
    UserMention(
      elem.getAs[Long]("id"),
      elem.getAs[String]("is_str"),
      elem.getAs[Array[Long]]("indices"),
      elem.getAs[String]("name"),
      elem.getAs[String]("screen_name"))
  }
}

mentions.first

我试图打电话时出现例外情况mentions.first:

scala>     mentions.first
15/06/23 22:15:06 ERROR Executor: Exception in task 0.0 in stage 5.0 (TID 8)
scala.MatchError: [List([187356243,187356243,List(3, 16),Lino Bocchini,linobocchini], [111123176,111123176,List(79, 95),Jean Wyllys,jeanwyllys_real])] (of class org.apache.spark.sql.catalyst.expressions.GenericRowWithSchema)
    at $line37.$read$$iwC$$iwC$$iwC$$iwC$$iwC$$iwC$$iwC$$iwC$$iwC$$iwC$$iwC$$iwC$$anonfun$1.apply(<console>:34)
    at $line37.$read$$iwC$$iwC$$iwC$$iwC$$iwC$$iwC$$iwC$$iwC$$iwC$$iwC$$iwC$$iwC$$anonfun$1.apply(<console>:34)
    at scala.Function1$$anonfun$andThen$1.apply(Function1.scala:55)
    at org.apache.spark.sql.catalyst.expressions.UserDefinedGenerator.eval(generators.scala:81)

这有什么不对？我知道它与类型有关但我还不知道它.

作为附加上下文,自动映射的结构是:

scala> mentions.printSchema
root
 |-- user_mentions: array (nullable = true)
 |    |-- element: struct (containsNull = true)
 |    |    |-- id: long (nullable = true)
 |    |    |-- id_str: string (nullable = true)
 |    |    |-- indices: array (nullable = true)
 |    |    |    |-- element: long (containsNull = true)
 |    |    |-- name: string (nullable = true)
 |    |    |-- screen_name: string (nullable = true)

注1:我知道有可能使用它来解决这个问题,HiveQL但是一旦有很多动力,我想使用数据帧.

SELECT explode(entities.user_mentions) as mentions
FROM tweets

注2:该UDF val isEmpty = udf((value: List[Any]) => value.isEmpty)是一个丑陋的黑客,我在这里失去了一些东西,但我想出了避免NPE的唯一途径

Answer 1

这是一个有效的解决方案，只需一个小技巧即可。

主要思想是通过声明 List[String] 而不是 List[Row] 来解决类型问题：

val mentions = tweets.explode("entities.user_mentions", "mention"){m: List[String] => m}

这将创建第二个名为“mention”、类型为“Struct”的列：

|            entities|             mention| 
+--------------------+--------------------+ 
|[List(),List(),Li...|[187356243,187356...| 
|[List(),List(),Li...|[111123176,111123...| 

现在执行一个map()来提取mention中的字段。getStruct(1) 调用获取每行第 1 列中的值：

case class Mention(id: Long, id_str: String, indices: Seq[Int], name: String, screen_name: String)
val mentionsRdd = mentions.map(
  row => 
    {  
      val mention = row.getStruct(1)
      Mention(mention.getLong(0), mention.getString(1), mention.getSeq[Int](2), mention.getString(3), mention.getString(4))
    }
)

并将 RDD 转换回 DataFrame：

val mentionsDf = mentionsRdd.toDF()

就这样吧！

|       id|   id_str|     indices|         name|    screen_name|
+---------+---------+------------+-------------+---------------+
|187356243|187356243| List(3, 16)|Lino Bocchini|   linobocchini|
|111123176|111123176|List(79, 95)|  Jean Wyllys|jeanwyllys_real|