Solution preview: removing punctuation is fastest and simplest with the built-in str.translate() and a translation table built from string.punctuation. For full Unicode punctuation, use unicodedata or a Unicode-aware regex.

Method 1: Delete ASCII punctuation with str.translate (fast, built-in)

This approach runs in C under the hood and is typically the most efficient for ASCII punctuation. See the documentation for details.

Step 1: Import string.

import string

Step 2: Build a translation table that deletes punctuation.

table = str.maketrans('', '', string.punctuation)

Step 3: Apply the table with translate.

s = "Hello, world! Python 3.12—fun?"
clean = s.translate(table)
print(clean)  # Hello world Python 312—fun

Step 4: Optionally replace punctuation with spaces instead of deleting.

space_table = str.maketrans({ch: ' ' for ch in string.punctuation})
spaced = s.translate(space_table)
normalized = ' '.join(spaced.split())
print(normalized)  # Hello world Python 3.12—fun

Notes:

  • string.punctuation covers ASCII punctuation only. It does not include punctuation like “ ” — 。 ！. See the reference.
  • To handle non-ASCII punctuation, use Method 3 or Method 4.

Option 2: Remove ASCII punctuation with re.sub

Regular expressions are concise and flexible. Escape the punctuation class once, then substitute. See re.sub in the documentation.

Step 1: Import re and string.

import re, string

Step 2: Compile a pattern that matches any ASCII punctuation.

pattern = re.compile(r'[%s]' % re.escape(string.punctuation))

Step 3: Substitute matches with empty strings (or a space).

s = "A test: regex-only, please!"
clean = pattern.sub('', s)
print(clean)  # A test regexonly please

Tip: \w includes letters, digits, and underscore, and \s matches whitespace; both are described in the documentation. If you prefer a whitelist, you can keep word and space characters: re.sub(r'[^\w\s]', '', s).

Approach 3: Remove all Unicode punctuation with unicodedata (built-in)

This approach removes any character whose Unicode category begins with 'P' (punctuation), not just ASCII. See the reference.

Step 1: Import unicodedata and sys.

import unicodedata, sys

Step 2: Build a deletion map for all code points in the Unicode range whose category starts with 'P'.

delete_punct = dict.fromkeys(
    i for i in range(sys.maxunicode + 1)
    if unicodedata.category(chr(i)).startswith('P')
)

Step 3: Translate the string using the map.

s = "Unicode: 「quotes」 — dashes… 你好，世界！"
clean = s.translate(delete_punct)
print(clean)  # Unicode  quotes  dashes  你好世界

Tip: If you also want to drop symbols like currency signs, extend the filter to include category 'S'.

Way 4: Use the third‑party “regex” module for Unicode properties

Python’s built-in re does not support \p{...} Unicode properties. The regex package supports them and can target punctuation precisely using \p{P}. Install it from the package page.

Step 1: Install the package.

pip install regex

Step 2: Import and compile a Unicode property pattern.

import regex
pattern = regex.compile(r'\p{P}+')

Step 3: Substitute punctuation with an empty string or a space.

s = "Mix: ASCII, Unicode… and 「symbols」!"
clean = pattern.sub('', s)
print(clean)  # Mix ASCII Unicode and symbols

Tip: To remove punctuation and symbols together, use r'[\p{P}\p{S}]+'.

Path 5: Quick comprehension/filter (simple, slower)

This pure-Python option is easy to read for small inputs, but is slower than the methods above.

Step 1: Import string.

import string

Step 2: Keep only non-punctuation characters.

s = "Keep it simple, okay?"
clean = ''.join(ch for ch in s if ch not in string.punctuation)
print(clean)  # Keep it simple okay

Note: This also relies on ASCII-only string.punctuation.

Practical tips:

  • Decide whether to delete punctuation or replace it with spaces; replacing then normalizing whitespace keeps word boundaries intact.
  • When using regex, remember \w includes underscore; if underscores should be removed, target them explicitly.
  • For very large texts or performance-critical code, prefer str.translate() with a prebuilt table.

That’s it—use str.translate for speed on ASCII, unicodedata or a Unicode-aware regex when you need to cover all punctuation across languages.