How to Remove Specific Words from a String in Python

To remove specific words from a string in Python, use re.sub() with word boundaries so only whole words go, then tidy the spaces. Plain replace() works too, but it also cuts the word out of longer words:

import re
re.sub(r"\bsale\b", "", "Big sale on sales tables").split()  # ['Big', 'on', 'sales', 'tables']

In this tutorial, we’ll compare four ways to do it, and I ran every script on Python 3.12.5 and pandas 3.0.6. Save each one as a .py file, open Command Prompt in that folder and run python file_name.py.

Remove a word with str.replace() and watch for partial matches

replace(old, "") deletes every occurrence of the text, wherever it sits. That’s fine for a unique word, but short words hide inside others:

text = "This listing is for a condo in Austin and it is pet friendly."

print(text.replace("is", ""))       # also hits "This" and "listing"
print(text.replace(" is", ""))      # safer here, still not bulletproof
print(text.replace(" is", "", 1))   # only the first match

Output:

Th lting  for a condo in Austin and it  pet friendly.
This listing for a condo in Austin and it pet friendly.
This listing for a condo in Austin and it is pet friendly.

The optional third argument limits how many matches change. If you’re removing arbitrary text rather than words, my guide on removing a substring from a string in Python covers slicing and the other options.

How do I remove only whole words from a string?

Wrap the words in \b, the regex word boundary, so free can’t match inside Freestyle. re.escape() keeps characters like . or $ literal, and re.IGNORECASE catches any capitalization:

import re

subject = "URGENT: Act now for your free upgrade on Freestyle plans in Austin"
banned = ["urgent", "free", "act now"]

pattern = r"\b(?:" + "|".join(map(re.escape, banned)) + r")\b"
cleaned = re.sub(pattern, "", subject, flags=re.IGNORECASE)
print(repr(cleaned))

cleaned = re.sub(r"\s{2,}", " ", cleaned).strip(" :")   # tidy the gaps
print(repr(cleaned))

Output:

':  for your  upgrade on Freestyle plans in Austin'
'for your upgrade on Freestyle plans in Austin'

Removing words always leaves extra spaces, so you’ll want the second pass. New to patterns? Start with regular expressions in Python, and see removing spaces from a string for more cleanup tricks.

Remove stopwords from a string with split() and join()

Stopwords are filler words like the, and or for. You split the text into words and keep the ones that aren’t in a set, with no regex needed. The catch is punctuation glued to a word:

import string

stop_words = {"the", "a", "an", "and", "of", "in", "to", "with"}
text = "Seattle has lots of cafes to work in. Pick one with fast Wi-Fi and a quiet corner."

naive = [w for w in text.split() if w.lower() not in stop_words]
print(" ".join(naive))

kept = [w for w in text.split()
        if w.strip(string.punctuation).lower() not in stop_words]
print(" ".join(kept))

Output:

Seattle has lots cafes work in. Pick one fast Wi-Fi quiet corner.
Seattle has lots cafes work Pick one fast Wi-Fi quiet corner.

A set makes each lookup quick even with hundreds of stopwords. For more on the split step, see splitting a sentence into a list of words, or strip it all first with removing punctuation from strings.

Remove words from a pandas column

For a whole column, Series.str.replace() takes the same pattern. Pass regex=True and case=False, then rejoin the words to squeeze the spaces:

import pandas as pd

products = pd.Series([
    "NEW Wireless Mouse - Sale",
    "Laptop Stand (sale) new arrival",
    "USB-C Hub Renewed",
])

pattern = r"\b(?:new|sale)\b"
clean = (products.str.replace(pattern, "", regex=True, case=False)
                 .str.replace(r"[()]|\s-\s", " ", regex=True)
                 .str.split().str.join(" "))
print(clean.to_string())

Output:

0          Wireless Mouse
1    Laptop Stand arrival
2       USB-C Hub Renewed

Need to know whether a word is there before you touch it? Use the checks in checking if a Python string contains a substring.

Word-removal methods compared

Method	Whole words only	Best for
s.replace(w, "")	No	Unique words, quick fixes
re.sub(r"\bw\b", "", s)	Yes	Lists of words, any case
split() + set + join()	Yes	Stopwords, long word lists
Series.str.replace(..., regex=True)	Yes	pandas columns

Keep going with these string cleanups:

• Remove prefixes from strings in Python
• Remove newlines from a string
• Count words in a file with Python
• Slice a string in Python

Frequently asked questions