spark-core-reduceByKey-01-order-count-for-each-day.scala

val orders = sc.textFile("/public/retail_db/orders")

val ordersMap = orders.
  map(o => (o.split(",")(1), 1))

ordersMap.
  reduceByKey((agg, ele) => agg + ele).
  take(10).
  foreach(println)

spark-core-reduceByKey-02-order-item-with-min-revenue-per-order-id.scala

val orderItems = sc.textFile("/public/retail_db/order_items")

val orderItemsMap = orderItems.
  map(oi => (oi.split(",")(1).toInt, oi.split(",")(4).toFloat))

orderItemsMap.
  reduceByKey((agg, ele) => agg.min(ele)).
  take(10).
  foreach(println)

spark-core-reduceByKey-03-revenue-and-count-per-order-id.scala

val orderItems = sc.textFile("/public/retail_db/order_items")

val orderItemsMap = orderItems.
  map(oi => (oi.split(",")(1).toInt, (oi.split(",")(4).toFloat, 1)))

orderItemsMap.
  reduceByKey((agg, ele) => (agg._1 + ele._1, agg._2 + ele._2)).
  take(10).
  foreach(println)