Question

我正在尝试创建一个可以在不更改原始数据帧的情况下更改数据框副本值的函数。这就是我到目前为止所做的：

def home_undervalued(df):
    local_df = df
    local_df['total_games'] = 0
    local_df['total_wins'] = 0
    cond_1 = local_df['predicted_spread'] > local_df['vegas_spread']
    cond_2 = local_df['actual_spread'] > local_df['vegas_spread']
    cond_3 = local_df['predicted_spread'] - local_df['vegas_spread'] >= 3

    local_df.loc[cond_1 & cond_3 , 'total_games'] = 1
    local_df.loc[cond_1 & cond_2 & cond_3 , 'total_wins'] = 1

    total_games = sum(local_df.total_games)
    total_wins = sum(local_df.total_wins)

    return float(total_wins) / float(total_games)

然后我用

调用该函数

home_undervalued(df)

它似乎有用，但后来我意识到df ['total_games']和df ['total_wins']的值已经改变了。我正在尝试更改local_df的值，但保留值df。关于如何解决这个问题的任何想法？

Answer 1

local_df = df只会创建对名为df的{{1}}的引用。如果你想创建一个完整的其他dataFrame（顺便提一下，我不推荐）你可以创建一个新的dataFrame作为local_df

Answer 2

只需使用DataFrame

的copy方法即可

local_df = df.copy(deep=True)

local_df中的这一行将被存储为df。

的副本

在不改变原始内容的情况下在函数内部复制数据帧

2 个答案: