Regex anchors: ^, $ and multiline mode explained

Most regex tokens match a character. Anchors match a position: the start of the text, the end, the gap between two characters. They take up no width at all.

They're also the most common thing missing from a validation pattern. This post covers ^ and $, what the m flag changes, and the corner cases that differ between languages. Every example is live, so edit the patterns and watch the lines change.

01A pattern without anchors checks less than you think

Regex engines search: they report a match if the pattern fits anywhere in the text. So a "five digit ZIP code" pattern without anchors is really "contains five digits in a row":

02^ and $

^ matches at the start of the text and $ at the end. Put them around the pattern and the whole input has to fit:

You can use just one of them. ^# means "starts with #", and \.pdf$ means "ends with .pdf":

03Anchors around alternation

| has the lowest priority of anything in a pattern, lower than the anchors. So ^cat|dog$ doesn't mean "exactly cat or dog". It means "starts with cat, or ends with dog":

Wrap the choices in a group so both anchors apply to both:

This one bites in real code, in allow-lists like ^admin|editor$ that quietly accept admin-old or not-an-editor.

04Multiline mode

By default, ^ and $ match only at the very start and end of the whole text, even if it contains line breaks. The m flag changes them to match at the start and end of every line:

const text = "apple\nbanana\ncherry";

text.match(/^\w+$/g);   // null: "apple\nbanana\ncherry" isn't one word
text.match(/^\w+$/gm);  // ["apple", "banana", "cherry"]

That's what you want when you're pulling things out of a file line by line. In Python it's re.MULTILINE, or (?m) at the start of the pattern:

import re

re.findall(r"^#+ .*$", markdown, re.MULTILINE)  # every heading line

The flag only changes ^ and $. A dot still won't cross a line break. That's a different flag, s, covered in the post on regex flags.

05The trailing newline trap

Here the languages split. In JavaScript, without m, $ matches only at the very end. In Python, $ also matches just before a newline at the very end:

import re

re.match(r"^\d{5}$", "90210\n")    # matches in Python
re.fullmatch(r"\d{5}", "90210\n")  # None

/^\d{5}$/.test("90210\n");  // false in JavaScript

So in Python, ^\d{5}$ accepts a ZIP code with a newline stuck on the end. That can matter if the value goes into a database or a file name. Two fixes:

• Use re.fullmatch(), which must consume the whole string.
• Use \A and \Z: in Python, \A is the very start and \Z the very end, with no newline exception. (Ruby, PCRE and Java spell the strict end \z.)

JavaScript has no \A or \z, but without the m flag its ^ and $ already behave like them.

06Word boundaries are anchors too

\b is another zero-width anchor: the position between a word character and a non-word character. It's how you match a whole word inside text:

07Try one

Catch the lines that are a hashtag and nothing else:

08Checklist

• Without anchors, a pattern only checks that the text contains a match.
• ^ is the start and $ is the end. Use both for validation.
• Group alternatives inside anchors: ^(cat|dog)$, not ^cat|dog$.
• The m flag makes ^ and $ work per line.
• In Python, $ allows one trailing newline. Use re.fullmatch() or \Z.
• \b anchors a whole word.

09Keep practising

• Anchor's away and Line wrangling are practice packs built around anchors.
• The guide's anchors section has the full list.