如何基于可变多列条件为Pandas DataFrame批量添加标签？

Got it, let's finish your labeling function! The key piece you're missing is combining multiple boolean conditions with logical AND (&) and using Pandas' .loc indexer to safely assign labels without warnings. Here's a step-by-step breakdown:

Step 1: Initialize the Label Column #

First, add a label column to your DataFrame with a default value (like 0 for rows that don't match any condition).

Step 2: Combine Conditions with Logical AND #

For each condition dictionary, you need to merge all column-value checks into a single boolean mask. We'll use functools.reduce to chain the & operator across all your conditions, ensuring only rows that meet all criteria get the label.

Step 3: Assign Labels Safely with .loc #

Use .loc to target rows matching the combined condition and set their label value. This avoids the common SettingWithCopyWarning that pops up with direct boolean indexing.

Complete Function Code #

import pandas as pd
from functools import reduce

def labeling(df, crit):
    # Initialize label column with default 0 for non-matching rows
    df['label'] = 0
    
    for dic in crit:
        # Extract the target label and make a copy of the condition dict to avoid modifying the original
        lbl = dic["label"]
        cond_dict = dic.copy()
        del cond_dict["label"]
        
        # Create a list of boolean conditions for each key-value pair
        conds = [df[key] == cond_dict[key] for key in cond_dict]
        
        # Combine all conditions using logical AND
        combined_condition = reduce(lambda x, y: x & y, conds)
        
        # Assign the label to rows that meet all conditions
        df.loc[combined_condition, 'label'] = lbl
    
    return df

Testing with Your Example #

Let's verify this works with your sample data and conditions:

Sample Input DataFrame

data = {
    'ip_src': ['192.168.84.129', '31.13.94.53', '192.168.1.101'],
    'ip_dst': ['192.168.84.128', '192.168.1.101', '31.13.94.53'],
    'ip_proto': [17.0, 17.0, 17.0],
    'frame_time_delta': [0.000000, 0.006656, 0.012948],
    'payload_size': [172.0, 176.0, 172.0],
    'src_port': [52165.0, 40002.0, 52165.0],
    'dst_port': [40002.0, 52165.0, 19305.0],
    'flow_dir': [1, 0, 1]
}
df = pd.DataFrame(data)

Conditions List

l_crit = [
    {"ip_src": "192.168.84.129", "ip_dst": "192.168.84.128", "label": 1},
    {"ip_src": "192.168.1.100", "ip_dst": "192.168.1.105", "dst_port": 9999, "label": 1},
    {"ip_src": "192.168.1.101", "ip_dst": "104.44.195.76", "label": 2},
    {"ip_src": "192.168.1.101", "ip_dst": "31.13.94.53", "ip_proto": 17, "label": 3},
    {"ip_src": "192.168.1.101", "dst_port": 19305, "label": 4}
]

Run the Function

labeled_df = labeling(df, l_crit)
print(labeled_df)

Output (Matches Your Expected Result)

ip_src         ip_dst  ip_proto  frame_time_delta  payload_size  src_port  dst_port  flow_dir  label
0  192.168.84.129  192.168.84.128      17.0           0.000000         172.0   52165.0   40002.0         1      1
1   31.13.94.53   192.168.1.101      17.0           0.006656         176.0   40002.0   52165.0         0      0
2  192.168.1.101    31.13.94.53      17.0           0.012948         172.0   52165.0   19305.0         1      4

Key Notes #

• We use dic.copy() to preserve your original condition dictionaries (so you can reuse them later if needed).
• reduce(lambda x, y: x & y, conds) ensures only rows that satisfy every condition in the dictionary get the label.
• .loc guarantees we're modifying the original DataFrame safely, avoiding common Pandas warnings.

内容的提问来源于stack exchange，提问作者sooaran