DataFrame多列内容比对及人性化可读格式转换技术咨询

Got it, let's break down your two pandas problems with practical, easy-to-follow solutions—these are super common when working with messy or unstructured tabular data!

1. Compare Multiple Columns in a DataFrame & Output Boolean Results to a New DataFrame #

If you need to check equality (or other conditions) between pairs of columns and store those True/False results in a separate DataFrame, pandas makes this straightforward.

Example Scenario #

Suppose your original DataFrame looks like this:

import pandas as pd

original_df = pd.DataFrame({
    'Product': ['Laptop', 'Phone', 'Tablet', 'Headphones'],
    'Price_2023': [999, 699, 299, 199],
    'Price_2024': [999, 749, 299, 249]
})

Solution Code #

To compare Price_2023 vs Price_2024 and store the boolean results in a new DataFrame:

# Create a new DataFrame for boolean comparisons
boolean_results = pd.DataFrame()

# Add comparison columns—you can extend this to any number of column pairs
boolean_results['Price_Unchanged'] = original_df['Price_2023'] == original_df['Price_2024']
boolean_results['Price_Increased'] = original_df['Price_2024'] > original_df['Price_2023']

# If you want to keep the original index or add reference columns, merge them together
boolean_results = pd.concat([original_df[['Product']], boolean_results], axis=1)

Customization Tip #

If you need more complex comparisons (like partial string matches), use pandas' string methods:

# Example: Check if Product name contains "Phone" (case-insensitive)
boolean_results['Is_Phone'] = original_df['Product'].str.contains('Phone', case=False)

2. Restructure a DataFrame for Human Readability (Combine Duplicate Rows) #

Your second task is to fix a DataFrame where the same question appears across multiple rows with different categories. We can use groupby() to collapse these duplicates into a single row per question, with all related categories grouped together.

Example Scenario #

Your original DF1 looks like this:

df1 = pd.DataFrame({
    'column1': ['question 1', 'question 2', 'question 1'],
    'column2': ['category A', 'category A', 'category B'],
    'column3': ['subcategory A', 'subcategory B', 'subcategory C']
})

Solution Code #

We'll group by the question column and aggregate the category columns into either comma-separated strings (for quick reading) or lists (for further processing):

# Option 1: Combine categories into comma-separated strings (great for human scanning)
readable_df = df1.groupby('column1').agg({
    'column2': lambda x: ', '.join(x.unique()),  # Use unique() to avoid duplicate categories
    'column3': lambda x: ', '.join(x.unique())
}).reset_index()

# Option 2: Combine categories into lists (better for downstream data operations)
readable_df_lists = df1.groupby('column1').agg({
    'column2': list,
    'column3': list
}).reset_index()

Result Explanation #

After running Option 1, your readable DataFrame will look like this:

This format is way easier to scan at a glance!

内容的提问来源于stack exchange，提问作者ben