re (web, native) – Edge Python

re is regular expressions on a backtracking engine. Import it by bare name, both runtimes resolve it with no manifest. To pin a different version, import it by URL or through a packages.json alias, see Modules.

Functions are match, search, fullmatch, findall, groups, span, and sub, all taking (pattern, string). compile(pattern) returns a pattern object with the same operations as methods. Flags go inline: (?i), (?s), (?m).

The match functions return the matched string or None, there are no Match objects. groups returns the capture groups as a list, span returns [start, end] in codepoint offsets, and findall with more than one group returns a list of lists. In sub, \1 and \g<name> expand groups. The syntax covers ., Unicode-aware classes \d \w \s and their negations, sets and ranges, anchors ^ $ \b \B, quantifiers * + ? {m,n} with lazy forms, capturing, non-capturing, and named groups, backreferences \1 and (?P=name), alternation, lookahead, and fixed-width lookbehind. Not supported: Unicode property classes \p{...}, \N{...} escapes, atomic groups, possessive quantifiers, conditionals, and scoped flags (?i:...). Invalid patterns raise ValueError. A per-match step budget raises RuntimeError on catastrophic backtracking instead of hanging. compile and the module-level functions share one compiled-pattern cache, so a hot loop parses each pattern once.

from re import search, sub, findall print(search(r'(\d+)-(\d+)', 'order 12-34')) print(sub(r'\s+', '_', 'a b c')) print(findall(r'\w+', 'one two three'))

Output · expected


12-34
a_b_c
['one', 'two', 'three']

groups and span give the captures and the offsets, and fullmatch returns None when the whole string does not match:

from re import groups, span, fullmatch print(groups(r'(\w+)@(\w+)', 'mail ada@edge')) print(span(r'\d+', 'abc 42 def')) print(fullmatch(r'\d+', '123')) print(fullmatch(r'\d+', '12a'))

Output · expected


['ada', 'edge']
[4, 6]
123
None

sub expands backreferences, and flags go inline:

from re import sub, search print(sub(r'(\w+) (\w+)', r'\2 \1', 'first last')) print(sub(r'(?P<y>\d{4})', r'[\g<y>]', 'in 2024')) print(search(r'(?i)EDGE', 'the edge engine'))

Output · expected


last first
in [2024]
edge

compile returns a pattern object with the same operations as methods:

from re import compile word = compile(r'\w+') print(word.findall('one two')) print(word.sub('#', 'one two three'))

Output · expected


['one', 'two']
# # #