spark sql 把字符串类型改为数值类型 并按照此列进行排序

只需要 order by CAST(cc.XX as INT)
XX为排序列名

val data4 = Seq(
      ("ming", "20190101004050", "man"),
      ("min", "20190101004050", "man"),
      ("mi", "20190101004050", "man"),
      ("kun", "20190102014050", "woman"),
      ("kun", "20190102012250", "man"),
      ("tian", "20190201025050", "man"),
      ("yarn", "20190205235050", "woman"),
      ("raa", "20190205113250", "man"),
      ("raa", "20190205115050", "woman"),
      ("kun", "20190201235050", "woman")
    ).toDF("name", "t", "gender")

    data4.show()

    data4.createOrReplaceTempView(s"data0")

    val hourdata = Seq(
      ("0"),("1"),("2"),("3"), ("4"),("5"),
      ("6"),("7"),("8"),("9"),("10"),("11"),("12"),
      ("13"),("14"),("15"),("16"),("17"),("18"),
      ("19"),("20"),("21"),("22"),("23")
    ).toDF("hour")
    hourdata.createOrReplaceTempView(s"hourdata")
    hourdata.show()

            val frame = spark.sql(
              s"select ”2019“ as dayId, * from "+
                s"("+
              s"select * from"+
                s" (select  " +
                s"count(name) as ${name}_cnt, count(distinct $name) as ${name}_distinct_cnt,"+
                s"count(gender) as ${gender}_cnt, count(distinct $gender) as ${gender}_distinct_cnt,"+
                s"aa.hour as hourId from"+
                s"(select name, gender" +
                s"hour(transformDateNew(t)) as hour " +
                s"from data0) aa group by aa.hour" +
                s") bb " +
                s" full outer join hourdata h on bb.hourId=h.hour"+
                s") cc  order by CAST(cc.hour as INT)"+
                s"").drop("hourId").na.fill(0)

结果：