dremio12.kt

val dailyHosts = days.groupBy("month", "day").count()

val totalReqPerDay = df.select(
    month(col("time")).alias("month"),
    dayofmonth(col("time")).alias("day")
).groupBy("month", "day").count()

val grouping = JavaConversions.asScalaBuffer(listOf("month", "day")).toSeq()
totalReqPerDay.join(dailyHosts, grouping)
    .select(
        col("month"),
        col("day"),
        totalReqPerDay.col("count").divide(dailyHosts.col("count"))
            .alias("avg")
    ).sort(col("avg").desc()).show()