根据值列表过滤火花数据帧中的条件

Question

我正在尝试根据值列表从数据帧中进行筛选，并且我能够按照示例1中给出的方式运行它。然而，当我将元素转换为list，然后将list传递到filter函数内的"isin"函数时，它不起作用(如示例2所示(。

val df1 = sc.parallelize(Seq((1,"abcd"), (2,"defg"), (3, "ghij"),(4,"xyzz"),(5,"lmnop"),(6,"pqrst"),(7,"wxyz"),(8,"lmnoa"),(9,"jklm"))).toDF("c1","c2")
//example 1:
val df2 = df1.filter(substring(col("c2"), 0, 3).isin("abc","def","ghi"))
//example 2:
val given_list = List("abc","def","ghi")
val df3 = df1.filter(substring(col("c2"), 0, 3).isin(given_list))

运行示例2时的错误消息如下所示：

Exception in thread "main" java.lang.RuntimeException: Unsupported literal type class scala.collection.immutable.$colon$colon List(abc, def, ghi)                                                  
19/10/22 17:03:10 INFO spark.SparkContext: Invoking stop() from shutdown hook                                                                                                                                   
19/10/22 17:03:10 INFO server.AbstractConnector: Stopped Spark@5817c15f{HTTP/1.1,[http/1.1]}{0.0.0.0:4040}                                                                                                      
19/10/22 17:03:10 INFO ui.SparkUI: Stopped Spark web UI at http://192.---.---.---:----                                                                                                                          
19/10/22 17:03:10 INFO spark.MapOutputTrackerMasterEndpoint: MapOutputTrackerMasterEndpoint stopped!                                                                                                            
19/10/22 17:03:10 INFO memory.MemoryStore: MemoryStore cleared                                                                                                                                                  
19/10/22 17:03:10 INFO storage.BlockManager: BlockManager stopped                                                                                                                                               
19/10/22 17:03:10 INFO storage.BlockManagerMaster: BlockManagerMaster stopped                                                                                                                                   
19/10/22 17:03:10 INFO scheduler.OutputCommitCoordinator$OutputCommitCoordinatorEndpoint: OutputCommitCoordinator stopped!                                                                                      
19/10/22 17:03:10 INFO spark.SparkContext: Successfully stopped SparkContext                                                                                                                                    
19/10/22 17:03:10 INFO util.ShutdownHookManager: Shutdown hook called                                                                                                                                           
19/10/22 17:03:10 INFO util.ShutdownHookManager: Deleting directory /tmp/spark-d17c100d-3e95-4016-a2fd-4e1e02b2449f          

提前感谢。

Answer 1

方法isin采用Any*varargs参数，而不是像List那样的集合。您可以使用"splat"运算符(即_*(，如下所示：

df1.filter(substring(col("c2"), 0, 3).isin(given_list: _*))

Spark 2.4+确实提供了采用Iterable集合的方法isInCollection，该方法可以如下使用：

df1.filter(substring(col("c2"), 0, 3).isInCollection(given_list))