根据循环内另一列的值将列的值更改为nan

Posted 2023-03-12

【中文标题】根据循环内另一列的值将列的值更改为nan

【英文标题】：Change the value of a column into a nan based on the value of another column inside a loop

【发布时间】：2020-10-12 10:03:58

【问题描述】：

我有大量带有后缀“mean”或“sum”的列。有时带有“平均”后缀的是NaN。发生这种情况时，我也想将带有“sum”后缀的那个也变成 NaN。我有大量变量，所以我需要 (?) 使用循环。我创建了一个假数据框，并添加了基于 SO 中类似帖子尝试过的 3 件事。不幸的是，没有任何效果

original_data_set = (pd.DataFrame
(

    'customerId':[1,2]
    ,'usage_1_sum':[100, 200]
    ,'usage_1_mean':[np.nan,100]
    ,'usage_2_sum':[420,330]
    ,'usage_2_mean':[45,np.nan]

)
             )

print('original dataset')
original_data_set

desired_data_set = (pd.DataFrame
(

    'customerId':[1,2]
    ,'usage_1_sum':[np.nan, 200]
    ,'usage_1_mean':[np.nan,100]
    ,'usage_2_sum':[420,np.nan]
    ,'usage_2_mean':[45,np.nan]

)
             )

print('desired dataset')
desired_data_set



holder_set = original_data_set.copy()

for number in range(1,3):
    holder_set['usage__sum'.format(number)] = (
        
        holder_set['usage__sum'.format(number)]
        .where(holder_set['usage__mean'.format(number)] == np.nan, np.nan
              )
                                                )

print('using an np.where statement changed all sum variables into NaN with no discretion')
holder_set


holder_set = original_data_set.copy()

for number in range(1,3):
    conditions = [holder_set['usage__mean'.format(number)]==np.nan]
    outcome = [np.nan]
    holder_set['usage__sum'.format(number)] = np.select(conditions, outcome, default=holder_set['usage__sum'.format(number)])
    
    
print('using an np.select did not have any effect on the dataframe')
holder_set


holder_set = original_data_set.copy()

for number in range(1,3):
    holder_set.loc[holder_set['usage__mean'.format(number)]==np.nan, 'usage__sum'.format(number)] = 12

print('using a loc did not have any effect on the dataframe')
holder_set

【问题讨论】：

也许可以尝试查看DataFrame.where() 功能。您应该能够直接索引到问题区域，而无需自己编写 for 循环。

【参考方案1】：

假设original 数据框为df：

df = pd.DataFrame('customerId': [1, 2], 'usage_1_sum': [100, 200], 'usage_1_mean': [
                  np.nan, 100], 'usage_2_sum': [420, 330], 'usage_2_mean': [45, np.nan])

使用Series.str.endswith 过滤以_mean 结尾的列，然后对于以_mean 结尾的列中的每一列，将_sum 列中的相应值更改为NaN，其中均值列中的值为@ 987654330@:

for col in df.columns[df.columns.str.endswith('_mean')]:
    df.loc[df[col].isna(), col.rstrip('_mean') + '_sum'] = np.nan

结果：

# print(df)
   customerId  usage_1_sum  usage_1_mean  usage_2_sum  usage_2_mean
0           1          NaN           NaN        420.0          45.0
1           2        200.0         100.0          NaN           NaN