如何用Pandas或SQL方式统计玩家累计积分并找出最高积分玩家

完全可以用Pandas原生方法和SQL实现，两种方案的性能都远高于你现在用iterrows遍历的实现，尤其数据量较大时优势更明显。

实现方案 #

Pandas 原生实现 #

核心思路是把winner_id和loser_id拆成独立的玩家行，保留对应积分后分组汇总：

# 将winner和loser字段打平为玩家ID列
player_points = matches_df.melt(
    id_vars=["points"],
    value_vars=["winner_id", "loser_id"],
    value_name="player_id"
)
# 按玩家ID分组求和累计积分
total_points = player_points.groupby("player_id")["points"].sum()
# 获取总积分最高的玩家ID
max_player_id = total_points.idxmax()

如果要同时拿到最高积分的数值，直接取total_points.max()即可。

SQL 实现 #

核心思路是用UNION ALL把获胜者、落败者的ID和对应积分合并后分组统计：

SELECT player_id, SUM(points) AS total_points
FROM (
    SELECT winner_id AS player_id, points FROM matches
    UNION ALL
    SELECT loser_id AS player_id, points FROM matches
) AS all_player_records
GROUP BY player_id
ORDER BY total_points DESC
LIMIT 1;

内容的提问来源于stack exchange，提问作者jebaseelan ravi