如何在DataFrame行中返回字典匹配值对应的多个键（以|分隔）

Got it, let's fix this up so you can capture all matching keys per row and join them with a |! Here's a straightforward approach using pandas and a custom function:

Step 1: Define Your Target Vocabulary Dictionary #

First, lay out the key-value pairs you're looking to extract. Let's use a category-based example that aligns with your sample product names:

# Replace this with your actual target vocabulary and mappings
target_map = {
    'Red': 'warm_color',
    'Blue': 'cool_color',
    'Azure': 'cool_color',
    'Ruby': 'warm_color',
    'Lace': 'fabric_detail',
    'Sweater': 'top_garment'
}

Step 2: Build a Custom Matching Function #

This function will scan each string, collect all matching values from your dictionary, and join them with |. If no matches are found, it returns np.nan (you can swap this for an empty string if preferred):

def extract_all_matches(text):
    # Gather all values where the corresponding key exists in the input text
    matched_values = [value for key, value in target_map.items() if key in text]
    # Join matches with | if any exist, else return NaN
    return '|'.join(matched_values) if matched_values else np.nan

Step 3: Apply the Function to Your DataFrame #

Use pandas' apply() method to run this function across every row in your target column:

import pandas as pd
import numpy as np

# Your sample DataFrame
df = pd.DataFrame({
    'Name': [
        'Red and Blue Lace Midi Dress',
        'Long Armed Sweater Azure and Ruby',
        'High Waisted Cotton Pants'  # Added a test row with no matches
    ]
})

# Add a new column with the combined matches
df['Matched_Categories'] = df['Name'].apply(extract_all_matches)

# Print the result
print(df)

Sample Output #

Name               Matched_Categories
0        Red and Blue Lace Midi Dress  warm_color|cool_color|fabric_detail
1  Long Armed Sweater Azure and Ruby  warm_color|cool_color|top_garment
2         High Waisted Cotton Pants                                NaN

Bonus: Whole-Word Matching (Avoid Partial Hits) #

If you want to prevent partial matches (e.g., not picking up "Red" in "Reddish"), use regex with word boundaries:

import re
def extract_all_matches(text):
    matched_values = []
    for key, value in target_map.items():
        # Use regex to match whole words only
        if re.search(rf'\b{re.escape(key)}\b', text):
            matched_values.append(value)
    return '|'.join(matched_values) if matched_values else np.nan

内容的提问来源于stack exchange，提问作者Bob Harris