代码如下：

python

import pandas as pd
import os
from datetime import datetime

# 读取数据
for file in os.listdir():
    if file.endswith('xlsx'):
        df = pd.read_excel(file, sheet_name='Tracing report',header=0)  # 假设输入文件名为 input.xlsx

# 处理合并单元格：填充空白值
df['ATD'] = df['ATD'].ffill()

# 将时间列转换为 datetime 类型
df['ATD'] = pd.to_datetime(df['ATD'],errors='coerce')  # E列是 ATD
df['ATA_T'] = pd.to_datetime(df.iloc[:,19],errors='coerce')  # T列是 ATA

df['ATA  KZH border'] = pd.to_datetime(df['ATA  KZH border'],errors='coerce')  # M列是 口岸

df = df.dropna(subset=['ATD', 'ATA_T','ATA  KZH border'])
# 计算时间差
df['整段时效'] = (df['ATA_T'] - df['ATD']).dt.days -1 # 转换为天
df['欧洲段时效'] = (df['ATA_T'] - df['ATA  KZH border']).dt.days -1  # 转换为天

# 按车站、月份和口岸分组
df['Month'] = df['ATD'].dt.to_period('M')  # 提取月份
# df['Month'] = df['ATD'].dt.month 
grouped = df.groupby(['Part No.', 'Month', 'ATA  CN border'], as_index=False)  # I列是 车站, K列是 口岸
# 计算每组的平均值
result = grouped.agg({
    '整段时效': lambda x: round(x.mean(), 2),
    '欧洲段时效': lambda x: round(x.mean(), 2)
})

result = result.rename(columns={
    'Part No.': '车站',
    'Month': '月份',
    'ATA  CN border': '口岸'
})
# 将结果保存到 Excel 文件
result.to_excel('output.xlsx', index=False)