如何按向上取整规则筛选指定行并创建新df？求助

Hey Dan, let's work through this problem together to get your desired DataFrame sorted out!

Step 1: Clarify the rounding logic #

First, we need to calculate the target C values based on your rounding rules for B:

• Round B up to the nearest multiple of 5: Use np.ceil(B / 5) * 5 (this gives 1135.0 in your example)
• Round B up to the nearest multiple of 10: Use np.ceil(B / 10) * 10 (this gives 1140.0 in your example)

Step 2: Implement priority-based filtering #

We'll first check for rows where C matches the 5-rounded target. If that returns an empty result (which is your current scenario), we'll fall back to the 10-rounded target. Here's a practical pandas/numpy implementation:

import pandas as pd
import numpy as np

# Replace this with your actual original DataFrame
original_df = pd.DataFrame({
    'B': [1132, 1137],
    'C': [1140.0, 1140.0]
})

# Calculate the two target values for each row
target_5 = np.ceil(original_df['B'] / 5) * 5
target_10 = np.ceil(original_df['B'] / 10) * 10

# First attempt: filter rows where C matches the 5-rounded target
new_df = original_df[original_df['C'] == target_5]

# If no rows are found, switch to the 10-rounded target
if new_df.empty:
    new_df = original_df[original_df['C'] == target_10]

print(new_df)

Quick notes to avoid issues #

• Double-check that columns B and C are numeric types (use original_df.dtypes to verify)
• If you're working with a single specific value of B (not an entire column), calculate the targets directly with that value instead of using the column
• This code handles row-wise checks—if you need to filter based on a global target (e.g., round one B value once and find all matching C rows), just adjust the target calculation to use that single value

内容的提问来源于stack exchange，提问作者Dan_99