spark在满足条件的列中获得最小值

我认为你的意思是你想循环数据框并有一个带退出的子循环。我找不到这样的例子，事实上我不确定它是否符合SPARK范式。获得相同的结果，处理更少：

import org.apache.spark.sql.functions._
import spark.implicits._
import org.apache.spark.sql.expressions.Window

val df = Seq((0, true), (1, true), (2,false), (3, true), (4,true), (5,true), (6,false), (7,false), (8,true), (9,false)).toDF("id","flag")
@transient val w1 = org.apache.spark.sql.expressions.Window.orderBy("id1")

val ids = df.where("flag = false")

        .select($"id".as("id1"))  

val ids2 = ids.select($"*", lag("id1",1,-1).over(w1).alias("prev_id"))
val ids3 = ids2.withColumn("prev_id1", col("prev_id")+1).drop("prev_id")

// Less and better performance at scale, this is better theoretically for Catalyst to bound partitions? Less work to do in any event.
// Some understanding of data required! And no grouping and min.
val withNextFalse = df.join(ids3, df("id") >= ids3("prev_id1") && df("id") <= ids3("id1"))

                 .select($"id", $"flag", $"id1".alias("nextOrCurrentFalse"))
                 .orderBy(asc("id"),asc("id"))

withNextFalse.show(false)
返回：