Question

我想编写一个泛型函数来用平均值或中值替换Null值或删除表，

def fill_null(df,col_name,action):
    df_tmp = df.apply(df[col_name].action())
    return df_tmp
if __name__ == '__main__':
    df_sal  =  pd.read_excel("sample_sal.xlsx")
    df = fill_null(df_sal,"Salary","mean")
    print(df)

错误获取：

AttributeError: 'Series' object has no attribute 'action'

我可以通过以下方式获取输出，但这不是req

if action == "mean":
  df[col_name].fillna(df[col_name].mean(), inplace=True)
  return df
elif action == "median":
  df[col_name].fillna(df[col_name].median(), inplace=True)
  return df
elif action == "drop":
  df = df.dropna(subset = [col_name])
  return df

Answer 1

我认为需要使用aggregate或apply来处理聚合函数：

def fill_null(df,col_name,action):
    df[col_name].fillna(df[col_name].agg(action), inplace=True)
    #alternative   
    #df[col_name].fillna(df[col_name].apply(action), inplace=True)
    return df

print (fill_null(df, 'D', 'mean'))

但如果还需要dropna：

def fill_null(df,col_name,action):

    if action == "dropna":
        df = df.dropna(subset = [col_name])
    else:
        df[col_name].fillna(df[col_name].agg(action), inplace=True)
    return df

如何在pandas中传入内置函数作为参数？

1 个答案: