在不在 DataFrame 中创建辅助列的情况下按日期字段执行合并

Question

在python pandas中是以下DataFrames:

|      date                   |     counter      |
|-----------------------------|------------------|
|  2022-01-01 10:00:02+00:00  |         34       |
|  2022-01-03 11:03:02+00:00  |         23       |
|  2022-02-01 12:00:05+00:00  |         12       |
|  2022-03-01 21:04:02+00:00  |         7        |

|      date                   |     holiday      |
|-----------------------------|------------------|
|  2022-01-01                 |       True       |
|  2022-01-02                 |       False      |
|  2022-01-03                 |       True       |
|  2022-02-01                 |       True       |
|  2022-02-02                 |       True       |
|  2022-02-03                 |       True       |
|  2022-03-01                 |       False      |
|  2022-03-02                 |       True       |
|  2022-03-03                 |       False      |

考虑到我不想创建带有日期的辅助列，我该如何合并两个 DataFrame？

|      date                   |     counter      |   holiday    |
|-----------------------------|------------------|--------------|
|  2022-01-01 10:00:02+00:00  |         34       |    True      |
|  2022-01-03 11:03:02+00:00  |         23       |    True      |
|  2022-02-01 12:00:05+00:00  |         12       |    True      |
|  2022-03-01 21:04:02+00:00  |         7        |    False     |

提前感谢您的帮助。

Answer 1

使用 Series.map with datetimes without times by Series.dt.normalize - 然后在 df2 输出中不创建辅助列：

df2['holiday'] = df2['date'].dt.normalize().map(df1.set_index('date')['holiday'])

merge_asof, but for avoid error need remove timezones by Series.dt.tz_convert的另一个想法：

df = pd.merge_asof(df1.assign(date = df1['date'].dt.tz_convert(None)).sort_values('date'), 
                   df2, on='date')
print (df)
                 date  counter  holiday
0 2022-01-01 10:00:02       34     True
1 2022-01-03 11:03:02       23     True
2 2022-02-01 12:00:05       12     True
3 2022-03-01 21:04:02        7    False

在不在 DataFrame 中创建辅助列的情况下按日期字段执行合并

Perform a merge by date field without creating an auxiliary column in the DataFrame

python

datetime

dataframe

pandas