如何用Pandas实现当列值大于阈值时删除后续指定行数？

Hey there! Let's work through this step by step to get your code working correctly for both your original and updated requirements.

First, Let's Address Your Initial Code Issue #

You wanted to delete the 3 rows after any row where Pout > 3, but your original code has a few key problems:

• You're looping through the values of Pout instead of their indices, so you can't actually locate which row triggered the condition.
• df.drop(df[3:]) deletes all rows starting at index 3, not the 3 rows after the matching row.
• Pandas' drop() returns a new DataFrame by default—you need to use inplace=True or reassign the result to save changes.

Corrected Code for Original Requirement (Delete 3 Rows After Pout > 3) #

import pandas as pd

df = pd.read_csv(file, sep=',', usecols=['Iin', 'Iout','Pout'])
print("Original Pout:\n", df['Pout'])

n_rows_to_delete = 3
# Get all indices where Pout > 3
condition_indices = df[df['Pout'] > 3].index.tolist()

# Process indices in reverse to avoid index shifts messing up later deletions
for idx in reversed(condition_indices):
    start = idx + 1
    end = idx + n_rows_to_delete
    # Don't go beyond the last row of the DataFrame
    end = min(end, df.index[-1])
    if start <= end:
        df.drop(range(start, end + 1), inplace=True)

print("\nPout after deletion:\n", df['Pout'])

Updating for Your New Requirement (Delete 5 Rows After c > 3) #

Your updated code sets values to None instead of deleting rows, and the break stops after the first matching row—let's fix that to handle all matching rows properly:

Optimized Code for Updated Requirement #

import pandas as pd

df = pd.DataFrame({
    'a': [1,5.0,1,2.3,2.1,2,1,3,4,7],
    'b': [1,4,0.2,4.5,8.2,1,2,3,4,7],
    'c': [1,4.5,5.4,6,2,4,2,3,4,7]
})

n_rows_to_delete = 5
# Get all indices where column 'c' is greater than 3
condition_indices = df[df['c'] > 3].index.tolist()

# Reverse the indices so deleting later rows doesn't affect earlier index positions
for idx in reversed(condition_indices):
    start_delete = idx + 1
    end_delete = idx + n_rows_to_delete
    # Ensure we don't try to delete rows that don't exist
    end_delete = min(end_delete, df.index[-1])
    if start_delete <= end_delete:
        # Delete the range of rows
        df.drop(range(start_delete, end_delete + 1), inplace=True)

print("Updated column 'c' after deletion:\n", df['c'])

Key Notes on This Approach #

• Reverse Index Processing: By iterating over the condition indices in reverse order, we avoid issues where deleting rows shifts the indices of remaining rows—this ensures we target the correct rows every time.
• Boundary Check: The min(end_delete, df.index[-1]) makes sure we don't try to delete rows beyond the end of the DataFrame if there aren't 5 rows left after a matching entry.
• Batch Deletion: Instead of modifying rows one by one (like your at calls), we delete full ranges at once, which is much more efficient for larger datasets.

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