如何使用python有效地填充“缺失时间模式”和“填充它们”具有特定值？

Posted 2023-04-18

【中文标题】如何使用python有效地填充“缺失时间模式”和“填充它们”具有特定值？

【英文标题】：How to fill the "missing time pattern" and "fill them" with particular value efficiently with python?

【发布时间】：2019-02-15 17:56:49

【问题描述】：

我想从以下位置“扩展”我的行：

+-------------+---------+-------+-------+
| Week Number | Weekday | Time  | Speed |
+-------------+---------+-------+-------+
|           1 | Monday  | 09.00 |     2 |
|           1 | Monday  | 12.00 |     2 |
|           1 | Monday  | 14.00 |     2 |
|           1 | Monday  | 15.00 |     1 |
|           1 | Tuesday | 08.00 |     4 |
|           1 | Tuesday | 10.00 |     2 |
|           1 | Tuesday | 11.00 |     3 |
|           1 | Tuesday | 13.00 |     2 |
+-------------+---------+-------+-------+

每天进入以下模式： 08.00, 09.00, 10.00, 11.00, 12.00, 13.00, 14.00, 15.00

+-------------+---------+-------+-------+
| Week Number | Weekday | Time  | Speed |
+-------------+---------+-------+-------+
|           1 | Monday  | 08.00 |     0 |
|           1 | Monday  | 09.00 |     2 |
|           1 | Monday  | 10.00 |     0 |
|           1 | Monday  | 11.00 |     0 |
|           1 | Monday  | 12.00 |     2 |
|           1 | Monday  | 13.00 |     0 |
|           1 | Monday  | 14.00 |     2 |
|           1 | Monday  | 15.00 |     1 |
|           1 | Tuesday | 08.00 |     4 |
|           1 | Tuesday | 09.00 |     0 |
|           1 | Tuesday | 10.00 |     2 |
|           1 | Tuesday | 11.00 |     3 |
|           1 | Tuesday | 12.00 |     0 |
|           1 | Tuesday | 13.00 |     3 |
|           1 | Tuesday | 14.00 |     0 |
|           1 | Tuesday | 15.00 |     0 |
+-------------+---------+-------+-------+

并用 0 填充“缺失”。 我该怎么办？

我正在使用带有 pandas 库的 python 3.6。

【问题讨论】：

time 列的 dtype 是什么？

timedelta64[ns]。有什么想法吗？

【参考方案1】：

import pandas as pd
df = pd.DataFrame('Week Number': 1, 'Weekday': ['Monday'] * 4 + ['Tuesday'] * 4, 'Time':['09.00', '12.00', '14.00', '15.00'] * 2,
                  'Speed': [2, 4] * 4)

假设 times、days 和 week_nums 都是扩展 DataFrame 的值

times = ['08.00', '09.00', '10.00', '11.00', '12.00', '13.00', '14.00', '15.00']
days = ['Monday', 'Tuesday']
week_nums = [1]

使用Speed = 0创建所有可能组合的DataFrame

from itertools import product
df_combinations = pd.DataFrame(list(product(, days, times, [0])), columns=['Week Number', 'Weekday', 'Time', 'Speed'])

连接两个数据框（df_combinations 必须是重复删除的第二个！）

df_new = pd.concat([df, df_combinations], ignore_index=True, sort=False)

创建重复的二进制掩码，删除它们并对数据帧进行排序

df_new = df_new[~df_new.duplicated(subset=['Week Number', 'Weekday', 'Time'], keep='first')]
df_new.sort_values(['Week Number', 'Weekday', 'Time'])