技术求助：将DataFrame转换为Array、Hashtable并按Manager迭代调整格式

First, let’s assume a sample original DataFrame structure (since you didn’t share the exact input format, this aligns with typical scenarios matching your requirements):

import pandas as pd

original_df = pd.DataFrame({
    'Manager': ['Alice', 'Alice', 'Bob', 'Bob', 'Bob', 'Charlie'],
    'Employee_Name': ['John', 'Jane', 'Mike', 'Sarah', 'Tom', 'Emma'],
    'Position': ['Software Engineer', 'Intern', 'Data Analyst', 'Intern', 'DevOps', 'UX Designer']
})

Step 1: Add Employment Type Column #

We’ll first populate the Employment_Type column using your rule: Intern roles are always Part-Time; all other positions are Full-Time. Using vectorized operations here is far more efficient than row-by-row loops:

import numpy as np

original_df['Employment_Type'] = np.where(original_df['Position'] == 'Intern', 'Part-Time', 'Full-Time')

Step 2: Iterate Per Manager and Transform to Target Format #

Below are two common approaches to handle per-Manager processing, depending on your exact mockup needs:

Approach 1: Process as Grouped DataFrames

If you need to work with each Manager’s data as a separate DataFrame (e.g., for exporting, reporting, or further calculations):

# Group the DataFrame by Manager
manager_groups = original_df.groupby('Manager')

# Iterate through each manager's group
for manager_name, group_df in manager_groups:
    print(f"=== Data for Manager: {manager_name} ===")
    # Adjust columns here to match your mockup's required fields
    display(group_df[['Employee_Name', 'Position', 'Employment_Type']])
    # Add custom logic here (e.g., save to CSV, generate summary stats)

Approach 2: Transform to Nested Structure (e.g., JSON-like dict)

If your mockup requires a nested format where each Manager maps to a list of their employees’ details:

result = {}
for manager_name, group_df in manager_groups:
    # Convert the group to a list of dictionaries matching your mockup structure
    result[manager_name] = group_df[['Employee_Name', 'Position', 'Employment_Type']].to_dict('records')

# Print the final structure
print(result)

Step 3: Customize to Match Your Exact Mockup #

If your target format has specific column names, ordering, or additional fields, tweak the code accordingly. For example, to rename columns and reorder them:

for manager_name, group_df in manager_groups:
    transformed = group_df.rename(columns={
        'Employee_Name': 'Full_Name',
        'Position': 'Job_Role'
    })[['Full_Name', 'Job_Role', 'Employment_Type']]
    print(f"=== Transformed Data for {manager_name} ===")
    print(transformed)

Key Tips #

• Vectorized operations (like np.where) are critical for performance with large datasets—avoid looping through individual rows if possible.
• The groupby method ensures clean separation of each Manager’s data, making iteration straightforward.
• If your original DataFrame has extra columns, simply adjust the subset selection (e.g., group_df[['Col1', 'Col2']]) to match your mockup’s required fields.

内容的提问来源于stack exchange，提问作者floader2022