Regular expressions

Here are more information about Python regular expressions:

• • The Python re module: http://docs.python.org/library/re.html

  • Regular expression HOWTO: http://docs.python.org/howto/regex.html

Use raw strings!

Always use Python raw strings. r"…", r'…', r"""…""", or r'''…'''

Regexp special characters

• • r"." matches any character

  • r"^…" / r"…$" matches the start / end of the string

  • r"…?" / r"…*" / r"…+" matches a regexp at most once / any number of times / at least once

  • r"…??" / r"…*?" / r"…+?" are the non-greedy alternatives

  • r"…|…|…" matches either this or that or that

  • r"[…]" matches any of the characters given

    • r"[^…]" matches any character not given

    • r"[^a-z]" matches any character between "a" and "z" (according to the unicode order)

    • r"[…-]" matches a minus (in addition to the other given characters)

    • r"[[…]" matches a closing bracket (in addition to the other given characters)

  • backslashed characters

    • r"\." / r"\[" / r"\\" / etc. matches the literal symbol

    • r"\s" matches a whitespace character

      • r"\S" matches non-whitespace

    • r"\w" matches a letter or a digit or underscore

      • r"\W" matches non-(letter|digit|underscore)

    • r"\d" matches a digit

      • r"\D" matches non-digit

    • r"\b" matches a word boundary

      • i.e., the empty string but only in the context of r"\W\w" or r"\w\W"

  • r"[^\W\d_]" matches only letters

    • try to understand why

    • the idea is taken from http://stackoverflow.com/questions/1673749

  • r"…(…)…" matches the whole regexp, but captures the part inside parenthesis so that you can look it up later

    • r"…(?:…)…" matches the regexp, and does not capture the parenthesis

Regexp functions

These are the main regexp functions:

• • re.compile(r"…", [flags])

  • returns an object with all the methods below (but with no pattern argument, since this is already compiled)

  • re.split(r"…", string, [maxsplit])

  • returns a list of strings, split by the pattern

    • if there are capturing parentheses r"…(…)…" in the pattern, then their values are also returned as part of the list

  • re.findall(r"…", string, [flags])

  • returns a list of all substrings that match the pattern

    • if there are n capturing parentheses r"…(…)…" in the pattern, then return a list of n-tuples

  • re.sub(r"…", replacement, string, [count])

  • substitutes each occurrence of the pattern with the replacement string

  • if there are n capturing parentheses r"…(…)…" in the pattern, then you can use \1, …, \n as backreferences in the replacement string

  • re.search(r"…", string, [flags])

    • returns a MatchObject, or None

    • read more about match objects here: http://docs.python.org/library/re.html#match-objects

  • re.finditer(r"…", string, [flags])

  • returns an iterator of MatchObject, that can be used in a for loop

Match objects

Match objects are returned by re.search and re.finditer. They have the following methods:

• • .group() => returns the matching string

  • m.group() == m.group(0)

  • .group(k) => returns the kth capture group (k=0 means the whole match)

  • .groups() => returns a tuple of all capture groups

  • .start() => returns the start position of the match in the search string

    • .start(k) => returns the start position of the kth capture group

  • .end() => returns the end position of the match in the search string

    • .end(k) => returns the end position of the kth capture group

  • .span() => returns a tuple (start, end)

  • m.span(k) == (m.start(k), m.end(k))