Pandas中DatetimeIndex无weekofmonth属性，月内周数独热编码失败求助

你已经成功实现了星期几和月份的独热编码：

import numpy as np
import pandas as pd

# 星期几独热编码
days=["Mon","Tue","Wed","Thu","Fri"]
for i in range(5):
    cv[days[i]] = (cv.index.dayofweek == i).astype(int)

# 月份独热编码
months = ["Jan", "Feb", "Mar", "Apr", "May", "Jun", "Jul", "Aug", "Sep", "Oct", "Nov", "Dec"]
for i in range(12):
    cv[months[i]] = (cv.index.month == i+1).astype(int)

但尝试生成月内周数独热编码时，因DatetimeIndex不存在weekofmonth属性，触发AttributeError。以下是两种可行的解决方法：

方法一：通过日期天数计算月内周数 #

直接利用日期的日数值计算所属月内周数，逻辑简单直观：

# 计算每个日期属于当月的第几周（每周按7天划分，1-7号为第1周，以此类推）
cv['week_of_month'] = (cv.index.day - 1) // 7 + 1

# 生成目标独热编码列
weeks = ['1st Week','2nd Week','3rd Week','4th Week']
for week_name in weeks:
    # 提取周数数字
    week_num = int(week_name.split()[0][0])
    cv[week_name] = (cv['week_of_month'] == week_num).astype(int)

# 若不需要中间列，可删除
# cv.drop('week_of_month', axis=1, inplace=True)

方法二：利用Period对象获取月内周数 #

借助Pandas的Period类型直接获取月内周数：

# 将DatetimeIndex转换为月度Period，提取月内周数
cv['week_of_month'] = cv.index.to_period('M').week

# 生成独热编码列
weeks = ['1st Week','2nd Week','3rd Week','4th Week']
for i in range(4):
    cv[weeks[i]] = (cv['week_of_month'] == i+1).astype(int)

# 可选删除中间计算列
# cv.drop('week_of_month', axis=1, inplace=True)

注意：部分月份可能存在第5周，若你的业务场景需要，可补充对应列；按你的需求保留前4周即可。

内容的提问来源于stack exchange，提问作者Tomward Matthias