Python按ID批量标记员工岗位与部门异动状态

原代码问题修正 #

你的原代码存在几处语法错误，先修正基础判断逻辑：

• 列索引错误：['Adjusted_Code'] 和 ['Adjusted Department'] 需改为 df['Adjusted_Code'] 和 df['Adjusted Department']
• 运算优先级问题：& 的优先级高于 !=，必须给每个比较条件单独加括号

修正后的单条记录判断代码：

import numpy as np
import pandas as pd

# 先标记单条记录是否满足同时换岗换部门的条件
df['is_transition'] = np.where(
    (df['Job_Code'] != df['Adjusted_Code']) & (df['Department'] != df['Adjusted Department']),
    'Yes',
    'No'
)

实现按ID批量标记需求 #

要达成只要ID存在至少一次符合条件的异动，该ID所有行的Transition都设为Yes的目标，用groupby+transform可以高效实现：

方法一：一步生成最终结果 #

无需临时列，直接按ID分组判断：

df['Transition'] = df.groupby('ID').transform(
    lambda group: 'Yes' if ((group['Job_Code'] != group['Adjusted_Code']) & (group['Department'] != group['Adjusted Department'])).any() else 'No'
)

方法二：基于临时列批量更新 #

如果已经生成了is_transition临时列，可简化为：

# 利用"Yes"字符串排序优先级高于"No"的特性，取每组最大值
df['Transition'] = df.groupby('ID')['is_transition'].transform('max')

# 或者用any()明确判断
df['Transition'] = df.groupby('ID')['is_transition'].transform(lambda x: 'Yes' if (x == 'Yes').any() else 'No')

最终输出示例 #

用你的测试数据运行后，结果会变为：

ID  Job_Code  Adjusted_Code   Job_Title  Adjusted_Job  Department Adjusted_Department Transition 
1    327          362         Associate    Manager       Sales         Sales             No
1    327          362         Associate    Manager       Sales         Sales             No
1    362          362         Associate    Manager       Sales         Sales             No
2    358          455         Consult      Advisor       HR            Finance           Yes
2    455          455         Advisor      Advisor       Finance       Finance           Yes
2    455          455         Advisor      Advisor       Finance       Finance           Yes            
3    215          381         Manager      Director      Tech          Sales             Yes
3    215          381         Manager      Director      Tech          Sales             Yes
3    215          381         Manager      Director      Tech          Sales             Yes
3    381          381         Manager      Director      Sales         Sales             Yes

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