Parse strings using a specification based on the Python format() syntax.
Latest commit 38e9fc8 Nov 17, 2014 @r1chardj0n3s Merge pull request #24 from lucien2k/master
Adding support for linux system log time


Parse strings using a specification based on the Python format() syntax.

parse() is the opposite of format()

The module is set up to only export parse(), search() and findall() when import * is used:

>>> from parse import *

From there it's a simple thing to parse a string:

>>> parse("It's {}, I love it!", "It's spam, I love it!")
<Result ('spam',) {}>
>>> _[0]

Or to search a string for some pattern:

>>> search('Age: {:d}\n', 'Name: Rufus\nAge: 42\nColor: red\n')
<Result (42,) {}>

Or find all the occurrances of some pattern in a string:

>>> ''.join(r.fixed[0] for r in findall(">{}<", "<p>the <b>bold</b> text</p>"))
'the bold text'

If you're going to use the same pattern to match lots of strings you can compile it once:

>>> from parse import compile
>>> p = compile("It's {}, I love it!")
>>> print(p)
<Parser "It's {}, I love it!">
>>> p.parse("It's spam, I love it!")
<Result ('spam',) {}>

("compile" is not exported for import * usage as it would override the built-in compile() function)

Format Syntax

A basic version of the Format String Syntax is supported with anonymous (fixed-position), named and formatted fields:

{[field name]:[format spec]}

Field names must be a valid Python identifiers, including dotted names; element indexes are supported (as they would make no sense.)

Numbered fields are also not supported: the result of parsing will include the parsed fields in the order they are parsed.

The conversion of fields to types other than strings is done based on the type in the format specification, which mirrors the format() behaviour. There are no "!" field conversions like format() has.

Some simple parse() format string examples:

>>> parse("Bring me a {}", "Bring me a shrubbery")
<Result ('shrubbery',) {}>
>>> r = parse("The {} who say {}", "The knights who say Ni!")
>>> print(r)
<Result ('knights', 'Ni!') {}>
>>> print(r.fixed)
('knights', 'Ni!')
>>> r = parse("Bring out the holy {item}", "Bring out the holy hand grenade")
>>> print(r)
<Result () {'item': 'hand grenade'}>
>>> print(r.named)
{'item': 'hand grenade'}
>>> print(r['item'])
hand grenade

Dotted names are possible though the application must make additional sense of the result:

>>> r = parse("Mmm, {food.type}, I love it!", "Mmm, spam, I love it!")
>>> print(r)
<Result () {'food.type': 'spam'}>
>>> print(r.named)
{'food.type': 'spam'}
>>> print(r['food.type'])

Format Specification

Most often a straight format-less {} will suffice where a more complex format specification might have been used.

Most of format()'s Format Specification Mini-Language is supported:


The differences between parse() and format() are:

  • The align operators will cause spaces (or specified fill character) to be stripped from the parsed value. The width is not enforced; it just indicates there may be whitespace or "0"s to strip.
  • Numeric parsing will automatically handle a "0b", "0o" or "0x" prefix. That is, the "#" format character is handled automatically by d, b, o and x formats. For "d" any will be accepted, but for the others the correct prefix must be present if at all.
  • Numeric sign is handled automatically.
  • The thousands separator is handled automatically if the "n" type is used.
  • The types supported are a slightly different mix to the format() types. Some format() types come directly over: "d", "n", "%", "f", "e", "b", "o" and "x". In addition some regular expression character group types "D", "w", "W", "s" and "S" are also available.
  • The "e" and "g" types are case-insensitive so there is not need for the "E" or "G" types.
Type Characters Matched Output
w Letters and underscore str
W Non-letter and underscore str
s Whitespace str
S Non-whitespace str
d Digits (effectively integer numbers) int
D Non-digit str
n Numbers with thousands separators (, or .) int
% Percentage (converted to value/100.0) float
f Fixed-point numbers float
e Floating-point numbers with exponent e.g. 1.1e-10, NAN (all case insensitive) float
g General number format (either d, f or e) float
b Binary numbers int
o Octal numbers int
x Hexadecimal numbers (lower and upper case) int
ti ISO 8601 format date/time e.g. 1972-01-20T10:21:36Z ("T" and "Z" optional) datetime
te RFC2822 e-mail format date/time e.g. Mon, 20 Jan 1972 10:21:36 +1000 datetime
tg Global (day/month) format date/time e.g. 20/1/1972 10:21:36 AM +1:00 datetime
ta US (month/day) format date/time e.g. 1/20/1972 10:21:36 PM +10:30 datetime
tc ctime() format date/time e.g. Sun Sep 16 01:03:52 1973 datetime
th HTTP log format date/time e.g. 21/Nov/2011:00:07:11 +0000 datetime
ts Linux system log format date/time e.g. Nov 9 03:37:44 datetime
tt Time e.g. 10:21:36 PM -5:30 time

Some examples of typed parsing with None returned if the typing does not match:

>>> parse('Our {:d} {:w} are...', 'Our 3 weapons are...')
<Result (3, 'weapons') {}>
>>> parse('Our {:d} {:w} are...', 'Our three weapons are...')
>>> parse('Meet at {:tg}', 'Meet at 1/2/2011 11:00 PM')
<Result (datetime.datetime(2011, 2, 1, 23, 0),) {}>

And messing about with alignment:

>>> parse('with {:>} herring', 'with     a herring')
<Result ('a',) {}>
>>> parse('spam {:^} spam', 'spam    lovely     spam')
<Result ('lovely',) {}>

Note that the "center" alignment does not test to make sure the value is centered - it just strips leading and trailing whitespace.

Some notes for the date and time types:

  • the presence of the time part is optional (including ISO 8601, starting at the "T"). A full datetime object will always be returned; the time will be set to 00:00:00. You may also specify a time without seconds.
  • when a seconds amount is present in the input fractions will be parsed to give microseconds.
  • except in ISO 8601 the day and month digits may be 0-padded.
  • the date separator for the tg and ta formats may be "-" or "/".
  • named months (abbreviations or full names) may be used in the ta and tg formats in place of numeric months.
  • as per RFC 2822 the e-mail format may omit the day (and comma), and the seconds but nothing else.
  • hours greater than 12 will be happily accepted.
  • the AM/PM are optional, and if PM is found then 12 hours will be added to the datetime object's hours amount - even if the hour is greater than 12 (for consistency.)
  • in ISO 8601 the "Z" (UTC) timezone part may be a numeric offset
  • timezones are specified as "+HH:MM" or "-HH:MM". The hour may be one or two digits (0-padded is OK.) Also, the ":" is optional.
  • the timezone is optional in all except the e-mail format (it defaults to UTC.)
  • named timezones are not handled yet.

Note: attempting to match too many datetime fields in a single parse() will currently result in a resource allocation issue. A TooManyFields exception will be raised in this instance. The current limit is about 15. It is hoped that this limit will be removed one day.

Result Objects

The result of a parse() operation is either None (no match) or a Result instance.

The Result instance has three attributes:

A tuple of the fixed-position, anonymous fields extracted from the input.
A dictionary of the named fields extracted from the input.
A dictionary mapping the names and fixed position indices matched to a 2-tuple slice range of where the match occurred in the input. The span does not include any stripped padding (alignment or width).

Custom Type Conversions

If you wish to have matched fields automatically converted to your own type you may pass in a dictionary of type conversion information to parse() and compile().

The converter will be passed the field string matched. Whatever it returns will be substituted in the Result instance for that field.

Your custom type conversions may override the builtin types if you supply one with the same identifier.

>>> def shouty(string):
...    return string.upper()
>>> parse('{:shouty} world', 'hello world', dict(shouty=shouty))
<Result ('HELLO',) {}>

If the type converter has the optional pattern attribute, it is used as regular expression for better pattern matching (instead of the default one).

>>> def parse_number(text):
...    return int(text)
>>> parse_number.pattern = r'\d+'
>>> parse('Answer: {number:Number}', 'Answer: 42', dict(Number=parse_number))
<Result () {'number': 42}>
>>> _ = parse('Answer: {:Number}', 'Answer: Alice', dict(Number=parse_number))
>>> assert _ is None, "MISMATCH"

You can also use the with_pattern(pattern) decorator to add this information to a type converter function:

>>> from parse import with_pattern
>>> @with_pattern(r'\d+')
... def parse_number(text):
...    return int(text)
>>> parse('Answer: {number:Number}', 'Answer: 42', dict(Number=parse_number))
<Result () {'number': 42}>

A more complete example of a custom type might be:

>>> yesno_mapping = {
...     "yes":  True,   "no":    False,
...     "on":   True,   "off":   False,
...     "true": True,   "false": False,
... }
... @with_pattern(r"|".join(yesno_mapping))
... def parse_yesno(text):
...     return yesno_mapping[text.lower()]

Version history (in brief):

  • 1.6.5 handle precision in float format (thanks Levi Kilcher)
  • 1.6.4 handle pipe "|" characters in parse string (thanks Martijn Pieters)
  • 1.6.3 handle repeated instances of named fields, fix bug in PM time overflow
  • 1.6.2 fix logging to use local, not root logger (thanks Necku)
  • 1.6.1 be more flexible regarding matched ISO datetimes and timezones in general, fix bug in timezones without ":" and improve docs
  • 1.6.0 add support for optional pattern attribute in user-defined types (thanks Jens Engel)
  • 1.5.3 fix handling of question marks
  • 1.5.2 fix type conversion error with dotted names (thanks Sebastian Thiel)
  • 1.5.1 implement handling of named datetime fields
  • 1.5 add handling of dotted field names (thanks Sebastian Thiel)
  • 1.4.1 fix parsing of "0" in int conversion (thanks James Rowe)
  • 1.4 add __getitem__ convenience access on Result.
  • 1.3.3 fix Python 2.5 issue.
  • 1.3.2 fix Python 3.2 issue.
  • 1.3.1 fix a couple of Python 3.2 compatibility issues.
  • 1.3 added search() and findall(); removed compile() from import * export as it overwrites builtin.
  • 1.2 added ability for custom and override type conversions to be provided; some cleanup
  • 1.1.9 to keep things simpler number sign is handled automatically; significant robustification in the face of edge-case input.
  • 1.1.8 allow "d" fields to have number base "0x" etc. prefixes; fix up some field type interactions after stress-testing the parser; implement "%" type.
  • 1.1.7 Python 3 compatibility tweaks (2.5 to 2.7 and 3.2 are supported).
  • 1.1.6 add "e" and "g" field types; removed redundant "h" and "X"; removed need for explicit "#".
  • 1.1.5 accept textual dates in more places; Result now holds match span positions.
  • 1.1.4 fixes to some int type conversion; implemented "=" alignment; added date/time parsing with a variety of formats handled.
  • 1.1.3 type conversion is automatic based on specified field types. Also added "f" and "n" types.
  • 1.1.2 refactored, added compile() and limited from parse import *
  • 1.1.1 documentation improvements
  • 1.1.0 implemented more of the Format Specification Mini-Language and removed the restriction on mixing fixed-position and named fields
  • 1.0.0 initial release

This code is copyright 2012-2013 Richard Jones <> See the end of the source file for the license of use.