Regular expressions

Pattern	Means	Matches
\d	a digit	7
\w	a letter, digit or _	a, 3, _
\s	whitespace	a space, a tab
.	any character at all	x
+	one or more	\d+ → 2026
*	zero or more
?	zero or one
{2,4}	between 2 and 4
[abc]	one of a, b, c
^ $	the start / the end of the line
(...)	a group that can be pulled out

\d{4}-\d{2}-\d{2} matches a date like 2026-09-20.

import re

log = "2026-09-20 13:05 ERROR user=42 message=timeout after 30s"

m = re.search(r"(\d{4}-\d{2}-\d{2}) (\d{2}:\d{2}) (\w+) user=(\d+)", log)
print(m.group(1), m.group(3), m.group(4))      # 2026-09-20 ERROR 42

# named groups are more readable
m = re.search(r"(?P<date>\d{4}-\d{2}-\d{2}).*user=(?P<uid>\d+)", log)
print(m.group("uid"))                          # 42

print(re.findall(r"\d+", log))                # ['2026','09','20','13','05','42','30']
print(re.sub(r"user=\d+", "user=***", log))   # mask the id

# raw strings: always use r"..." so that \d is not interpreted by Python

When a regex is the wrong tool: HTML and JSON (use a parser — nested structures cannot be matched with a regex), and patterns that grow longer than one line. A regex you cannot read in a month's time is a bug waiting to happen.