{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [],
   "source": [
    "import pyspark\n",
    "sc = pyspark.SparkContext.getOrCreate()\n",
    "\n",
    "spark = SparkSession \\\n",
    "    .builder \\\n",
    "    .appName(\"Spark SQL Finger 3\") \\\n",
    "    .getOrCreate()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [],
   "source": [
    "datos_jugadores_regtemp = [\n",
    "    (1, 'Manu Ginobili', 30, 35, 8, 8, 5, 2),\n",
    "    (2, 'Kobe Bryant', 40, 30, 3, 5, 20, 2),\n",
    "    (3, 'LeBron James', 58, 20, 3, 8, 30, 2),\n",
    "    (4, 'Andre Iguodala', 35, 15, 5, 10, 20, 4),\n",
    "    (5, 'Carmelo Anthony', 40, 10, 6, 6, 22, 1)\n",
    "]\n",
    "\n",
    "datos_jugadores_po = [\n",
    "    #1, 2, 3, 4, 5, 6 ,7 ,8 ,9\n",
    "    (1, 1, 1, 8, 3, 4, 2, 12, 1),\n",
    "    (1, 1, 3, 4, 1, 5, 2, 15, 1),\n",
    "    (1, 1, 4, 12, 2, 3, 2, 10, 1),\n",
    "    (2, 1, 1, 18, 3, 3, 3, 8, 2),\n",
    "    (2, 1, 4, 9, 3, 4, 1, 5, 1),\n",
    "    (3, 2, 1, 12, 6, 4, 1, 3, 1),\n",
    "    (3, 2, 2, 15, 8, 5, 3, 7, 1),\n",
    "    (3, 2, 3, 22, 6, 2, 1, 4, 2),\n",
    "    (3, 2, 4, 18, 10, 4, 2, 4, 1),\n",
    "    (4, 2, 1, 12, 2, 1, 1, 9, 0),\n",
    "    (4, 2, 2, 25, 8, 4, 2, 4, 3),\n",
    "    (4, 2, 4, 15, 4, 2, 2, 8, 1),\n",
    "    (1, 3, 3, 2, 1, 0, 0, 3, 2),\n",
    "    (5, 3, 1, 8, 1, 2, 0, 6, 0),\n",
    "    (5, 3, 1, 7, 0, 0, 1, 6, 1),\n",
    "]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [],
   "source": [
    "df_tr = spark.createDataFrame(datos_jugadores_regtemp)\n",
    "df_po = spark.createDataFrame(datos_jugadores_po)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "['_1', '_2', '_3', '_4', '_5', '_6', '_7', '_8']"
      ]
     },
     "execution_count": 16,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#  La columna del promedio de asistencias es _4.\n",
    "df_tr.columns"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "['_1', '_2', '_3', '_4', '_5', '_6', '_7', '_8', '_9']"
      ]
     },
     "execution_count": 17,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#  La columna con la cantidad de asistencias es _8.\n",
    "df_po.columns"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [],
   "source": [
    "df_tr.createOrReplaceTempView('temporada_regular')\n",
    "df_po.createOrReplaceTempView('playoffs')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [],
   "source": [
    "#  Creamos una tabla nueva agrupando por jugador y partido, para calcular el total de asistencias por partido.\n",
    "po_suma = spark.sql(\"SELECT _1, _2, sum(_8) as asistencias_totales FROM playoffs GROUP BY _1,_2\")\n",
    "po_suma.createOrReplaceTempView('po_asistencias')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [],
   "source": [
    "#  La columna _2 en \"temporada_regular\" es el nombre del jugador, _1 es el id de jugador en ambas tablas.\n",
    "prom_y_max_asistencias = spark.sql(\"SELECT _2 as nombre, _4 as prom_asistencias, max_asistencias FROM temporada_regular as tr\\\n",
    "          JOIN (SELECT _1, max(asistencias_totales) as max_asistencias FROM po_asistencias GROUP BY _1) as po\\\n",
    "                  ON tr._1 = po._1\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "+---------------+----------------+---------------+\n",
      "|         nombre|prom_asistencias|max_asistencias|\n",
      "+---------------+----------------+---------------+\n",
      "|Carmelo Anthony|              10|             12|\n",
      "|  Manu Ginobili|              35|             37|\n",
      "|   LeBron James|              20|             18|\n",
      "|    Kobe Bryant|              30|             13|\n",
      "| Andre Iguodala|              15|             21|\n",
      "+---------------+----------------+---------------+\n",
      "\n"
     ]
    }
   ],
   "source": [
    "prom_y_max_asistencias.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [],
   "source": [
    "prom_y_max_asistencias.createOrReplaceTempView('resultado')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [],
   "source": [
    "query_final = spark.sql(\"SELECT nombre FROM resultado as res WHERE res.max_asistencias > prom_asistencias\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "+---------------+\n",
      "|         nombre|\n",
      "+---------------+\n",
      "|Carmelo Anthony|\n",
      "|  Manu Ginobili|\n",
      "| Andre Iguodala|\n",
      "+---------------+\n",
      "\n"
     ]
    }
   ],
   "source": [
    "query_final.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "PySpark",
   "language": "python",
   "name": "pyspark"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.6.6"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}