pyspark - 如何找到过滤的行总和

Question

我有一个包含两列的数据集：Country、Adclicks。如何找到广告点击次数最多的国家/地区？

Country | Ad Click
USA       1
USA       0
USA       1
PR        0
PR        0
PR        1

Answer 1

假设您的 DataFrame 定义为变量“df”，则如下所示：

import pyspark.sql.functions as psf

# Get aggregate sum
s = df.groupby("Country").agg({'Ad Click': 'sum'})

# Get and display top country
s.registerTempTable("sums_table")
query = """
    SELECT Country
    FROM sums_table
    WHERE `sum(Ad Click)` = (
        SELECT MAX(`sum(Ad Click)`)
        FROM sums_table)
"""
top_country = spark.sql(query).collect()
print(top_country[0]["Country"])