如何用apply从DataFrame多列提取括号内内容

Got it, let's work through this. You mentioned extracting bracket content works perfectly with vectors, but hitting a wall when using apply on DataFrame columns—here's how to get this working reliably.

Step 1: Define a Robust Extraction Function #

First, let's create a function that handles all edge cases (like missing values, non-string entries, or text without brackets) so it doesn't break when applied to your DataFrame. We'll use regex for the extraction:

import re
import pandas as pd

def pull_bracket_text(input_text):
    # Handle NaNs, numbers, or non-string values first
    if pd.isna(input_text) or not isinstance(input_text, str):
        return pd.NA
    # Match content inside parentheses (adjust regex for other brackets!)
    # Use r'\[(.*?)\]' for square brackets, r'\{(.*?)\}' for curly braces
    matches = re.findall(r'\((.*?)\)', input_text)
    # Return first match if exists, else NA
    return matches[0] if matches else pd.NA

Step 2: Apply the Function to Multiple Columns #

You have two straightforward options depending on whether you want to process all columns or just a subset:

Option 1: Process Every Column in the DataFrame

Use applymap—it's designed for element-wise operations across the entire DataFrame, which is exactly what we need here:

# Sample DataFrame to test with
df = pd.DataFrame({
    'fruit': ['Apple (Red)', 'Banana (Yellow)', 'Grape'],
    'animal': ['Cat (Mammal)', pd.NA, 'Dog (Canine)'],
    'vehicle': ['Car (Sedan)', '', 'Bike (Mountain)']
})

# Apply extraction to all columns
df_extracted = df.applymap(pull_bracket_text)

Option 2: Process Only Specific Columns

If you don't want to touch every column, target the ones you need and use applymap on that subset:

# List columns you want to process
target_columns = ['fruit', 'animal']

# Overwrite the original columns with extracted content
df[target_columns] = df[target_columns].applymap(pull_bracket_text)

Common Pitfalls to Fix Why Your Original Attempt Failed #

• No handling for non-string values: If your DataFrame has NaNs, integers, or floats, your original vector-based code might not account for these—our function checks for this explicitly.
• Wrong apply axis: If you used df.apply(..., axis=1), that processes rows instead of individual elements. Stick to applymap for element-wise work, or use df[col].apply(pull_bracket_text) for single columns.
• Mismatched regex: Double-check your regex matches the bracket type you're targeting (e.g., square vs. round brackets). Our example uses round—adjust the pattern if needed.

内容的提问来源于stack exchange，提问作者Al14