如何用另一列内容替换句子任意位置的指定字符串

Got it, let's get this sorted so you can replace matching terms anywhere in your sentences, not just the first word. Here's a straightforward approach that works with your existing NLTK setup:

Step-by-Step Explanation #

1. Tokenize the Sentence: Split your input sentence into individual words using NLTK's word_tokenize—this lets us check every word, no matter its position.
2. Case-Insensitive Matching: Convert each token to lowercase to compare against your all-lowercase col1 values (so "Aaa1" matches "aaa1").
3. Replace Matching Tokens: Iterate through each token, replacing it with your target value (like "NNP") if its lowercase version exists in col1.
4. Reconstruct the Sentence: Join the processed tokens back into a single string.

Code Implementation #

from nltk import word_tokenize

def replace_any_position(sentence, target_lower_terms, replacement):
    # Split sentence into individual words
    tokens = word_tokenize(sentence)
    # Check each token: replace if lowercase matches any term in target list
    processed_tokens = [
        replacement if token.lower() in target_lower_terms else token
        for token in tokens
    ]
    # Join tokens back into a sentence
    return ' '.join(processed_tokens)

# Example usage
if __name__ == "__main__":
    # Your col1 values (all lowercase)
    col1 = ["aaa1"]
    # Test case 1: Target in middle of sentence
    input_sentence1 = "Hello Aaa1 my very good friend"
    output1 = replace_any_position(input_sentence1, col1, "NNP")
    print(output1)  # Output: Hello NNP my very good friend
    
    # Test case 2: Target at start of sentence
    input_sentence2 = "Aaa1 my very good friend"
    output2 = replace_any_position(input_sentence2, col1, "NNP")
    print(output2)  # Output: NNP my very good friend

Notes for DataFrame Integration #

If your col1 is a column in a pandas DataFrame, convert it to a set for faster lookups (sets have O(1) membership checks vs lists' O(n)):

import pandas as pd

# Example DataFrame
df = pd.DataFrame({
    "col1": ["aaa1", "bbb2"],
    "col2": ["Hello Aaa1 my friend", "Bbb2 is here"]
})

# Convert col1 to a set of lowercase terms
target_set = set(df["col1"].str.lower())

# Apply the function to each row in col2
df["col2_processed"] = df["col2"].apply(
    lambda x: replace_any_position(x, target_set, "NNP")
)

print(df["col2_processed"])

This will handle replacements anywhere in the sentence, exactly what you need!

内容的提问来源于stack exchange，提问作者belriso