数据源是:
col1
------
false
false
true
false
false
true
true
false
我添加了一个新列,如果col1的值为true,则col2的值增加1。 我希望:
col1,col2
--
false,0
false,0
true,1
false,1
false,1
true,2
true,3
false,3
如何添加呢?
答案 0 :(得分:0)
可以使用窗口功能:
val df = Seq(false, false, true, false, false, true, true, false).toDF("col1")
val ordered = df
.withColumn("id", monotonically_increasing_id())
.withColumn("increment", when($"col1" === true, 1).otherwise(0))
val idWindow = Window.orderBy("id").rowsBetween(Window.unboundedPreceding, Window.currentRow)
val result = ordered.select($"col1", sum($"increment").over(idWindow).alias("col2"))
输出:
+-----+----+
|col1 |col2|
+-----+----+
|false|0 |
|false|0 |
|true |1 |
|false|1 |
|false|1 |
|true |2 |
|true |3 |
|false|3 |
+-----+----+
答案 1 :(得分:0)
scala> import org.apache.spark.sql.expressions.Window
scala> val w = Window.partitionBy(lit(1)).orderBy(lit(1))
scala> val w1 = Window.partitionBy(lit(1)).orderBy("rn")
scala> df.withColumn("tmp", when($"col1" === true, 1).otherwise(0)).withColumn("rn", row_number.over(w)).withColumn("col2", sum("tmp").over(w1)).select("col1","col2").show
+-----+----+
| col1|col2|
+-----+----+
|false| 0|
|false| 0|
| true| 1|
|false| 1|
|false| 1|
| true| 2|
| true| 3|
|false| 3|
+-----+----+