我正在使用此JSON创建的数据帧:
{"id" : "1201", "name" : "satish", "age" : "25"},
{"id" : "1202", "name" : "krishna", "age" : "28"},
{"id" : "1203", "name" : "amith", "age" : "39"},
{"id" : "1204", "name" : "javed", "age" : "23"},
{"id" : "1205", "name" : "mendy", "age" : "25"},
{"id" : "1206", "name" : "rob", "age" : "24"},
{"id" : "1207", "name" : "prudvi", "age" : "23"}
最初看起来像这样的数据框:
+---+----+-------+
|age| id| name|
+---+----+-------+
| 25|1201| satish|
| 28|1202|krishna|
| 39|1203| amith|
| 23|1204| javed|
| 25|1205| mendy|
| 24|1206| rob|
| 23|1207| prudvi|
+---+----+-------+
我需要的是将所有年龄相同的学生分组,并根据他们的ID订购。到目前为止,这就是我接近的方式:
*注意:我很确定,与使用withColumn("newCol", ..)
添加新列相比,使用select("newCol")
是更有效的方法,但是我不知道如何更好地解决它
val conf = new SparkConf().setAppName("SimpleApp").set("spark.driver.allowMultipleContexts", "true").setMaster("local[*]")
val sc = new SparkContext(conf)
val sqlContext = new org.apache.spark.sql.SQLContext(sc)
import sqlContext.implicits._
val df = sqlContext.read.json("students.json")
import org.apache.spark.sql.functions._
import org.apache.spark.sql.expressions._
val mergedDF = df.withColumn("newCol", collect_list(struct("age","id","name")).over(Window.partitionBy("age").orderBy("id"))).select("List")
我得到的输出是:
[WrappedArray([25,1201,satish], [25,1205,mendy])]
[WrappedArray([24,1206,rob])]
[WrappedArray([23,1204,javed])]
[WrappedArray([23,1204,javed], [23,1207,prudvi])]
[WrappedArray([28,1202,krishna])]
[WrappedArray([39,1203,amith])]
现在,如何过滤有多个元素的行?也就是说,我希望我的最终数据框架是:
[WrappedArray([25,1201,satish], [25,1205,mendy])]
[WrappedArray([23,1204,javed], [23,1207,prudvi])]
到目前为止,我最好的方法是:
val mergedDF = df.withColumn("newCol", collect_list(struct("age","id","name")).over(Window.partitionBy("age").orderBy("id")))
val filterd = mergedDF.withColumn("count", count("age").over(Window.partitionBy("age"))).filter($"count" > 1).select("newCol")
但是我一定缺少一些东西,因为结果不是预期的:
[WrappedArray([23,1204,javed], [23,1207,prudvi])]
[WrappedArray([25,1201,satish])]
[WrappedArray([25,1201,satish], [25,1205,mendy])]
您可以使用size()
过滤数据:
import org.apache.spark.sql.functions.{col,size}
mergedDF.filter(size(col("newCol"))>1).show(false)
+---+----+------+-----------------------------------+
|age|id |name |newCol |
+---+----+------+-----------------------------------+
|23 |1207|prudvi|[[23,1204,javed], [23,1207,prudvi]]|
|25 |1205|mendy |[[25,1201,satish], [25,1205,mendy]]|
+---+----+------+-----------------------------------+