如何重复DataFrame行并为id追加索引以保持其唯一性？

Got it, let's fix this! You're already using np.repeat to duplicate rows, but the id column is ending up with duplicates. Here's a straightforward, efficient way to make sure each id stays unique by appending an index suffix:

Step 1: Start with your original DataFrame #

First, let's set up an example to work with (adjust this to match your actual data):

import pandas as pd
import numpy as np

# Sample DataFrame
df = pd.DataFrame({
    'id': ['user_1', 'user_2', 'user_3'],
    'score': [85, 92, 78]
})

Step 2: Duplicate rows with np.repeat #

Use np.repeat to create your duplicated rows (we'll repeat each row 3 times here—change the number to match your needs):

# Repeat each row 3 times
repeat_count = 3
repeated_df = df.loc[np.repeat(df.index, repeat_count)].reset_index(drop=True)

At this point, the id column will have duplicates (e.g., user_1 appears 3 times).

Step 3: Make id unique with group-based indexing #

Use groupby and cumcount() to add a unique suffix to each duplicated id. This method is efficient even for large DataFrames:

# Append a sequential number to each duplicated id
repeated_df['id'] = (
    repeated_df['id'] 
    + '_' 
    + (repeated_df.groupby('id').cumcount() + 1).astype(str)
)

What the output looks like #

Your final DataFrame will have unique ids while keeping all duplicated row data intact:

id  score
0  user_1_1     85
1  user_1_2     85
2  user_1_3     85
3  user_2_1     92
4  user_2_2     92
5  user_2_3     92
6  user_3_1     78
7  user_3_2     78
8  user_3_3     78

Customization tips #

• If you don't want an underscore, replace '_' with another separator (like '-') or remove it entirely to get user_11, user_12, etc.
• If your original id is numeric, you can modify the logic to add a decimal or offset instead (e.g., repeated_df['id'] = repeated_df['id'] + (repeated_df.groupby('id').cumcount() + 1)/10 to get 1.1, 1.2, etc.).

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