How-to April 26, 2026

Regex match anything except specific characters

The negated character class [^...] is the right tool — clearer than lazy quantifiers, faster than alternation.

The short answer

Use a negated character class — square brackets with a caret at the start:

[^abc]     match any single character except a, b, or c
[^"]       match any character except a double quote
[^\d]      match any non-digit character (same as \D)
[^A-Z]     match anything except uppercase letters

To match multiple characters, add a quantifier: [^"]*, [^abc]+, etc.

How negation works

Inside [...], a caret ^ at the very beginning means "match any character that is NOT one of these." Without the caret, it matches "any character that IS one of these."

[abc]     match a, b, or c
[^abc]    match anything except a, b, c

If the caret isn't the first character, it's literal: [a^b] matches a, ^, or b.

Common use cases

Content between delimiters (the right way)

"[^"]*"        content between double quotes
\[[^\]]*\]    content between square brackets
<[^>]+>        HTML tag with attributes

These are faster and clearer than the lazy alternatives ".*?", \[.*?\], <.+?>.

Words without specific characters

[a-zA-Z]+         only letters
[^aeiouAEIOU]+    only consonants (well, no vowels)
[a-zA-Z][a-zA-Z0-9_]*    identifier (letter then letters/digits/underscores)

Lines not starting with a comment

^[^#]                    line doesn't start with #
^[^#].*                  match the whole line if it doesn't start with #
^(?!#).*$                same idea with negative lookahead

Shorthand negations

Several built-in classes are already negations of common groups:

Shorthand	Equivalent	Meaning
`\D`	`[^0-9]`	non-digit
`\W`	`[^A-Za-z0-9_]`	non-word
`\S`	`[^\s]`	non-whitespace

Excluding multiple things

Combine literal characters, ranges, and shorthand classes inside the negation:

[^\d\s.,!?]      not a digit, whitespace, or punctuation
[^a-zA-Z0-9]      not alphanumeric
[^"']            not a quote (either kind)
[^<>]             not an angle bracket

Excluding newlines

. already excludes newlines by default. So . and [^\n] are similar — but watch the differences:

.            any char except \n (and \r in some flavors)
[^\n]        any char except \n only

For most use cases . is fine. Use [^\n] when you specifically want to allow \r (Windows line endings) but not \n.

Negative lookarounds for "not preceded by"

Character classes negate one character at a time. For "not preceded by X" or "not followed by X", use lookarounds:

(?<!@)word        match "word" if NOT preceded by @
word(?!ing)       match "word" if NOT followed by "ing"

Combining negation with other constraints

You can negate one class and intersect with another using lookarounds:

(?=[^aeiou])[a-z]     a lowercase letter that's not a vowel

Or, in flavors with class intersection (Java, ES2024 with /v flag):

[a-z&&[^aeiou]]      lowercase consonants (Java)
[[a-z]--[aeiou]]    same idea (ES2024 /v)

Common mistakes

Putting the caret in the wrong place

[a^bc]    matches a, ^, b, or c — caret is literal!
[^abc]    matches anything except a, b, c

The negation caret MUST be the first character inside the brackets.

Forgetting that ] needs escaping

Inside [...], a literal ] usually needs escaping (or being the first character right after the opening bracket / caret):

[]a-z]     matches ] or lowercase letter — ] as first char is literal
[\]a-z]    safer — escape it explicitly

The takeaway

The negated character class [^...] is one of the most useful constructs in regex. Use it for:

Content between delimiters (much cleaner than lazy)
"Anything except this single character or set"
Excluding specific characters from a broader pattern

For "not preceded by" or "not followed by" — patterns about position rather than a single character — use negative lookarounds instead.