我能做到这一点在Scala中,也许你能得到的我在做什么的依据和实现它自己:
scala> val df1=sc.parallelize(Seq((1001,2),(1002,3),(1003,4))).toDF("t1","v1")
df1: org.apache.spark.sql.DataFrame = [t1: int, v1: int]
scala> val df2=sc.parallelize(Seq((1001,3),(1002,4),(1005,4))).toDF("t2","v2")
df2: org.apache.spark.sql.DataFrame = [t2: int, v2: int]
scala> df1.join(df2,df1("t1")===df2("t2"))
res1: org.apache.spark.sql.DataFrame = [t1: int, v1: int ... 2 more fields]
scala> res1.show
+----+---+----+---+
| t1| v1| t2| v2|
+----+---+----+---+
|1002| 3|1002| 4|
|1001| 2|1001| 3|
+----+---+----+---+
scala> import org.apache.spark.sql.functions._
import org.apache.spark.sql.functions._
scala> val result=res1.withColumn("foo",res1("v1") * res1("v2"))
result: org.apache.spark.sql.DataFrame = [t1: int, v1: int ... 3 more fields]
scala> result.show
+----+---+----+---+---+
| t1| v1| t2| v2|foo|
+----+---+----+---+---+
|1002| 3|1002| 4| 12|
|1001| 2|1001| 3| 6|
+----+---+----+---+---+
我希望这能解决你的问题。
Thanx解决方案,我试着用这些概念,但没有得到确切的解决方案。时间戳列包含值如2016-09-01 15:31:58 + 00:00。我想加载csv并将其拆分成列,结果应该像(t1,v * v2)。 –
然后使用先将其转换为火花时间戳,然后执行这些步骤,或者如果您想以简单方式执行此操作,则只需使用字符串即可。 –