使用Pandas将每两行网球胜率数据合并为一行生成对阵数据集

前置准备 #

首先导入Pandas库，读取原始数据。如果你是从csv文件读取，直接替换为pd.read_csv("你的文件路径.csv")即可：

import pandas as pd

# 示例：构造原始DataFrame，和你给出的输入格式一致
raw_df = pd.DataFrame(
    data = [
        ["player1", "35%"],
        ["player2", "65%"],
        ["player3", "20%"],
        ["player4", "80%"],
        ["player5", "90%"],
        ["player6", "10%"]
    ],
    columns = ["player", "probability_of_win"]
)

数据转换实现 #

方案1：分组聚合（逻辑直观，适合需要额外处理分组数据的场景） #

核心逻辑是给每两行（同一场比赛）分配相同的分组ID，再按分组聚合出两个选手的信息：

# 按索引每两行分为一组，生成比赛ID
raw_df["match_id"] = raw_df.index // 2

# 按比赛ID聚合，分别取出每组的第1、2位选手信息
result_df = raw_df.groupby("match_id").agg(
    playerA = ("player", "first"),
    playerB = ("player", "last"),
    probability_of_win_A = ("probability_of_win", "first"),
    probability_of_win_B = ("probability_of_win", "last")
).reset_index(drop=True)

方案2：数组重塑（执行效率高，适合大数据量场景） #

核心逻辑是把原始二维数组按每两行一组重塑，直接生成目标结构的DataFrame：

# 把原始数据按每两行一组重塑，每行4个元素对应目标表的4列
reshaped_arr = raw_df.values.reshape(-1, 4)
# 生成结果DataFrame并调整列顺序
result_df = pd.DataFrame(
    reshaped_arr,
    columns = ["playerA", "probability_of_win_A", "playerB", "probability_of_win_B"]
)[["playerA", "playerB", "probability_of_win_A", "probability_of_win_B"]]

输出验证 #

打印result_df即可得到你需要的目标格式：

内容的提问来源于stack exchange，提问作者Khaled Koubaa