RDD和sparksql的求分组topN

RDD

16、计算出每个地区的地点人数最多的前3名？
studentRDD.map(line => (line._4,line._7.toInt))
.groupByKey()
.map(line => {
(line._1, line.2.toList.sortWith( > _).take(3)) //按照降序进行排列
}).collect().foreach(println)

sparkSql

/4.4、分别统计出男生和女生的分数前三名
// spark.sql(“SELECT *,Row_Number() OVER (PARTITION BY sex ORDER BY scores DESC) AS TOPN FROM student_scores”).show(100)
spark.sql(“select * from (SELECT *, Row_Number() OVER (partition by sex ORDER BY scores desc) topn FROM student_scores) t1 where t1.topn<=3”).show()