基于分类值将Pandas DataFrame多列多行合并为单行（Python）

Hey there! Let's work through this DataFrame reshaping problem you're stuck on in Python 3.4 with Pandas. Since you didn't share the exact structure of your model DataFrame or the precise desired output, I'll cover some common reshaping techniques that are likely to solve your issue, with concrete examples you can adapt.

Common Reshaping Techniques #

1. Pivot Rows to Columns #

If your goal is to convert categorical rows into distinct columns (e.g., grouping by an ID and spreading attributes across columns), pivot() is your go-to tool.

Sample Model Input:

import pandas as pd

model = pd.DataFrame({
    'UserID': [101, 101, 102, 102],
    'Metric': ['Age', 'Score', 'Age', 'Score'],
    'Value': [25, 88, 30, 92]
})

Code to Reshape:

# Pivot the DataFrame
reshaped_df = model.pivot(index='UserID', columns='Metric', values='Value').reset_index()
# Remove the auto-generated column name for cleanliness
reshaped_df.columns.name = None

print(reshaped_df)

Output:

UserID  Age  Score
0     101   25     88
1     102   30     92

2. Group and Aggregate Multiple Rows #

If you need to combine multiple rows for the same group into a single row (e.g., merging lists of values), use groupby() with custom aggregation functions.

Sample Model Input:

model = pd.DataFrame({
    'OrderID': [5001, 5001, 5002, 5002, 5002],
    'Product': ['Laptop', 'Mouse', 'Phone', 'Charger', 'Case'],
    'Price': [999, 25, 699, 30, 15]
})

Code to Reshape:

# Group by OrderID and aggregate products/prices into comma-separated strings
reshaped_df = model.groupby('OrderID').agg({
    'Product': lambda x: ', '.join(x),
    'Price': lambda x: ', '.join(map(str, x))
}).reset_index()

print(reshaped_df)

Output:

OrderID               Product     Price
0     5001        Laptop, Mouse  999, 25
1     5002  Phone, Charger, Case  699, 30, 15

3. Handle Complex Reshaping with melt (If Needed) #

If your desired output involves unpivoting columns into rows (the reverse of pivoting), use pd.melt(). This is useful if you need to flatten wide tables into long formats.

Next Steps #

If none of these methods match your specific use case, share a small sample of your model DataFrame (you can use model.head().to_dict() to get a readable format) and the exact desired output structure. That way, we can craft a solution tailored exactly to your data.

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