你可以使用pandas中的merge函数来实现。试试这个：

import pandas as pd

# Assuming df1 has columns 'start_date' and 'end_date', and df2 has 'timestamp'

merged_df = pd.merge_asof(df2.sort_values('timestamp'), df1, left_on='timestamp', right_on='start_date')

result_df = merged_df[(merged_df['timestamp'] >= merged_df['start_date']) & (merged_df['timestamp'] <= merged_df['end_date'])]

output_df = result_df[['Image']]

字符串
这应该给你给予一个基于你的条件的df2的过滤版本（当然）

我会使用numpy来进行比较（=>and<=）* 元素方式 *，然后只保留福尔斯在至少一个时间间隔内的时间戳（使用any）：

import numpy as np

# check my first edit if you need to parse the datetime columns

start, end = df1["date_start"].to_numpy(), df1["date_end"].to_numpy()
timestamps = df2["timestamp"].to_numpy()[:, None]

# does the timestamp falls between at least one interval [ST, END] ?
m = np.logical_and(timestamps >= start, timestamps <= end).any(axis=1)

out = df2.loc[m]

字符串
输出量：

print(out)

                  timestamp  value
0  2023-10-23T17-18-56-341Z   1500

[1 rows x 2 columns]

型