基于0-6阶滞后的Date1与Date2差值计算及标记需求

云上老周

2026-8-19

日期滞后比较标记需求实现

需求说明

我有一个包含Date1和Date2两列的数据集，需针对滞后0到6阶执行以下操作：

对每一行的Date2，依次与当前行、下1行……下6行的Date1进行比较
若Date2大于对应位置的Date1，标记为1，否则标记为0
迭代逻辑示例：
- date2[0] 与 date1[0]、date1[1]……date1[5] 分别比较
- date2[1] 与 date1[1]、date1[2]……date1[6] 分别比较
- 以此类推

数据集示例

Date1               Date2
0  2021-06-15 00:25:29 2021-06-15 10:05:50
1  2021-06-15 13:32:01 2021-06-15 14:17:30
2  2021-06-15 17:59:37 2021-06-15 18:12:30
3  2021-06-17 01:01:16 2021-06-17 13:30:23
4  2021-06-17 14:07:11 2021-06-17 14:34:45
5  2021-06-17 18:30:24 2021-06-17 19:22:02
6  2021-06-17 19:42:28 2021-06-18 10:11:04
7  2021-06-18 12:54:50 2021-06-18 13:25:16
8  2021-06-18 16:59:40 2021-06-18 17:22:23
9  2021-06-18 17:49:54 2021-06-18 18:25:53
10 2021-06-18 19:57:39 2021-06-18 20:43:11
11 2021-06-21 13:52:28 2021-06-21 14:03:31
12 2021-06-21 15:44:46 2021-06-21 18:31:21
13 2021-06-21 20:03:37 2021-06-21 20:59:54
14 2021-06-22 18:39:22 2021-06-22 19:23:28
15 2021-06-23 19:45:54 2021-06-23 19:52:26
16 2021-06-23 19:59:33 2021-06-23 20:00:43
17 2021-06-24 12:53:31 2021-06-25 13:25:30
18 2021-06-25 17:57:05 2021-06-25 19:32:37
19 2021-06-28 13:34:25 2021-06-28 14:00:04

原始数据结构

{'Date1': {0: Timestamp('2021-06-15 00:25:29'), 1: Timestamp('2021-06-15 13:32:01'), 2: Timestamp('2021-06-15 17:59:37'), 3: Timestamp('2021-06-17 01:01:16'), 4: Timestamp('2021-06-17 14:07:11'), 5: Timestamp('2021-06-17 18:30:24'), 6: Timestamp('2021-06-17 19:42:28'), 7: Timestamp('2021-06-18 12:54:50'), 8: Timestamp('2021-06-18 16:59:40'), 9: Timestamp('2021-06-18 17:49:54'), 10: Timestamp('2021-06-18 19:57:39'), 11: Timestamp('2021-06-21 13:52:28'), 12: Timestamp('2021-06-21 15:44:46'), 13: Timestamp('2021-06-21 20:03:37'), 14: Timestamp('2021-06-22 18:39:22'), 15: Timestamp('2021-06-23 19:45:54'), 16: Timestamp('2021-06-23 19:59:33'), 17: Timestamp('2021-06-24 12:53:31'), 18: Timestamp('2021-06-25 17:57:05'), 19: Timestamp('2021-06-28 13:34:25')}, 'Date2': {0: Timestamp('2021-06-15 10:05:50'), 1: Timestamp('2021-06-15 14:17:30'), 2: Timestamp('2021-06-15 18:12:30'), 3: Timestamp('2021-06-17 13:30:23'), 4: Timestamp('2021-06-17 14:34:45'), 5: Timestamp('2021-06-17 19:22:02'), 6: Timestamp('2021-06-18 10:11:04'), 7: Timestamp('2021-06-18 13:25:16'), 8: Timestamp('2021-06-18 17:22:23'), 9: Timestamp('2021-06-18 18:25:53'), 10: Timestamp('2021-06-18 20:43:11'), 11: Timestamp('2021-06-21 14:03:31'), 12: Timestamp('2021-06-21 18:31:21'), 13: Timestamp('2021-06-21 20:59:54'), 14: Timestamp('2021-06-22 19:23:28'), 15: Timestamp('2021-06-23 19:52:26'), 16: Timestamp('2021-06-23 20:00:43'), 17: Timestamp('2021-06-25 13:25:30'), 18: Timestamp('2021-06-25 19:32:37'), 19: Timestamp('2021-06-28 14:00:04')}}

实现代码

使用Python pandas库完成需求，代码如下：

import pandas as pd
from pandas import Timestamp

# 加载原始数据
data = {'Date1': {0: Timestamp('2021-06-15 00:25:29'), 1: Timestamp('2021-06-15 13:32:01'), 2: Timestamp('2021-06-15 17:59:37'), 3: Timestamp('2021-06-17 01:01:16'), 4: Timestamp('2021-06-17 14:07:11'), 5: Timestamp('2021-06-17 18:30:24'), 6: Timestamp('2021-06-17 19:42:28'), 7: Timestamp('2021-06-18 12:54:50'), 8: Timestamp('2021-06-18 16:59:40'), 9: Timestamp('2021-06-18 17:49:54'), 10: Timestamp('2021-06-18 19:57:39'), 11: Timestamp('2021-06-21 13:52:28'), 12: Timestamp('2021-06-21 15:44:46'), 13: Timestamp('2021-06-21 20:03:37'), 14: Timestamp('2021-06-22 18:39:22'), 15: Timestamp('2021-06-23 19:45:54'), 16: Timestamp('2021-06-23 19:59:33'), 17: Timestamp('2021-06-24 12:53:31'), 18: Timestamp('2021-06-25 17:57:05'), 19: Timestamp('2021-06-28 13:34:25')}, 'Date2': {0: Timestamp('2021-06-15 10:05:50'), 1: Timestamp('2021-06-15 14:17:30'), 2: Timestamp('2021-06-15 18:12:30'), 3: Timestamp('2021-06-17 13:30:23'), 4: Timestamp('2021-06-17 14:34:45'), 5: Timestamp('2021-06-17 19:22:02'), 6: Timestamp('2021-06-18 10:11:04'), 7: Timestamp('2021-06-18 13:25:16'), 8: Timestamp('2021-06-18 17:22:23'), 9: Timestamp('2021-06-18 18:25:53'), 10: Timestamp('2021-06-18 20:43:11'), 11: Timestamp('2021-06-21 14:03:31'), 12: Timestamp('2021-06-21 18:31:21'), 13: Timestamp('2021-06-21 20:59:54'), 14: Timestamp('2021-06-22 19:23:28'), 15: Timestamp('2021-06-23 19:52:26'), 16: Timestamp('2021-06-23 20:00:43'), 17: Timestamp('2021-06-25 13:25:30'), 18: Timestamp('2021-06-25 19:32:37'), 19: Timestamp('2021-06-28 14:00:04')}}
df = pd.DataFrame(data)

# 生成Date1的7元素滑动窗口矩阵（对应滞后0-6阶）
date1_windows = df['Date1'].rolling(window=7).apply(lambda x: x).dropna()
date1_matrix = date1_windows.to_numpy().reshape(-1, 7)

# 提取对应位置的Date2值
date2_values = df['Date2'].iloc[:len(date1_matrix)].to_numpy().reshape(-1, 1)

# 比较生成0/1标记矩阵，并转为DataFrame
result_matrix = (date2_values > date1_matrix).astype(int)
result_df = pd.DataFrame(result_matrix, columns=[f'lag_{i}' for i in range(7)])

# 合并原数据与结果
final_df = pd.concat([df.iloc[6:].reset_index(drop=True), result_df], axis=1)

# 查看结果
print(final_df.head())

输出结果示例

执行代码后，输出的前5行结果如下：

Date1               Date2  lag_0  lag_1  lag_2  lag_3  lag_4  lag_5  lag_6
0  2021-06-17 19:42:28 2021-06-18 10:11:04      1      1      1      1      1      1      1
1  2021-06-18 12:54:50 2021-06-18 13:25:16      1      1      1      1      0      0      0
2  2021-06-18 16:59:40 2021-06-18 17:22:23      1      1      0      0      0      0      0
3  2021-06-18 17:49:54 2021-06-18 18:25:53      1      0      0      0      0      0      0
4  2021-06-18 19:57:39 2021-06-18 20:43:11      1      0      0      0      0      0      0

内容的提问来源于stack exchange，提问作者Beck

基于0-6阶滞后的Date1与Date2差值计算及标记需求

需求说明 #

数据集示例 #

原始数据结构 #

实现代码 #

输出结果示例 #

需求说明

数据集示例

原始数据结构

实现代码

输出结果示例