from pyspark.sql import SparkSession
from pyspark.sql.functions import explode, split, col

def spark_word_count():
    spark = SparkSession.builder.appName("spark_word_count").getOrCreate()
    path = "file:///mnt/data/word_count.txt"
    df = spark.read.text(path)
    df_wc = df.select(explode(split('value', ' ')).alias('word'))
    df_top100 = df_wc.groupBy('word').count().sort('count', ascending=False).limit(100)
    df_top100.show()
    spark.stop()

if __name__ == '__main__':
    spark_word_count()