Pyparsing CSV string with random quotes

Question

I have a string like the following:

<118>date=2010-05-09,time=16:41:27,device_id=FE-2KA3F09000049,log_id=0400147717,log_part=00,type=statistics,subtype=n/a,pri=information,session_id=o49CedRc021772,from="[email protected]",mailer="mta",client_name="example.org,[194.177.17.24]",resolved=OK,to="[email protected]",direction="in",message_length=6832079,virus="",disposition="Accept",classifier="Not,Spam",subject="=?windows-1255?B?Rlc6IEZ3OiDg5fDp5fog+fno5fog7Pf46eHp7S3u4+Tp7SE=?="

I tried using CSV module and it didn't fit, cause i haven't found a way to ignore what's quoted. Pyparsing looked like a better answer but i haven't found a way to declare all the grammars.

Currently, i am using my old Perl script to parse it, but i want this written in Python. if you need my Perl snippet i will be glad to provide it.

Any help is appreciated.

Answer 1

It might be better to leverage an existing parser than to use ad-hoc regexs.

parse_http_list(s)
    Parse lists as described by RFC 2068 Section 2.

    In particular, parse comma-separated lists where the elements of
    the list may include quoted-strings.  A quoted-string could
    contain a comma.  A non-quoted string could have quotes in the
    middle.  Neither commas nor quotes count if they are escaped.
    Only double-quotes count, not single-quotes.

parse_keqv_list(l)
    Parse list of key=value strings where keys are not duplicated.

Example:

>>> pprint.pprint(urllib2.parse_keqv_list(urllib2.parse_http_list(s)))
{'<118>date': '2010-05-09',
 'classifier': 'Not,Spam',
 'client_name': 'example.org,[194.177.17.24]',
 'device_id': 'FE-2KA3F09000049',
 'direction': 'in',
 'disposition': 'Accept',
 'from': '[email protected]',
 'log_id': '0400147717',
 'log_part': '00',
 'mailer': 'mta',
 'message_length': '6832079',
 'pri': 'information',
 'resolved': 'OK',
 'session_id': 'o49CedRc021772',
 'subject':'=?windows-1255?B?Rlc6IEZ3OiDg5fDp5fog+fno5fog7Pf46eHp7S3u4+Tp7SE=?=',
 'subtype': 'n/a',
 'time': '16:41:27',
 'to': '[email protected]',
 'type': 'statistics',
 'virus': ''}

Answer 2

I'm not sure what you're really looking for, but

import re
data = "date=2010-05-09,time=16:41:27,device_id=FE-2KA3F09000049,log_id=0400147717,log_part=00,type=statistics,subtype=n/a,pri=information,session_id=o49CedRc021772,from=\"[email protected]\",mailer=\"mta\",client_name=\"example.org,[194.177.17.24]\",resolved=OK,to=\"[email protected]\",direction=\"in\",message_length=6832079,virus=\"\",disposition=\"Accept\",classifier=\"Not,Spam\",subject=\"=?windows-1255?B?Rlc6IEZ3OiDg5fDp5fog+fno5fog7Pf46eHp7S3u4+Tp7SE=?=\""
pattern = r"""(\w+)=((?:"(?:\\.|[^\\"])*"|'(?:\\.|[^\\'])*'|[^\\,"'])+)"""
print(re.findall(pattern, data))

gives you

[('date', '2010-05-09'), ('time', '16:41:27'), ('device_id', 'FE-2KA3F09000049'),
 ('log_id', '0400147717'), ('log_part', '00'), ('type', 'statistics'),
 ('subtype', 'n/a'), ('pri', 'information'), ('session_id', 'o49CedRc021772'),
 ('from', '"[email protected]"'), ('mailer', '"mta"'),
 ('client_name', '"example.org,[194.177.17.24]"'), ('resolved', 'OK'),
 ('to', '"[email protected]"'), ('direction', '"in"'),
 ('message_length', '6832079'), ('virus', '""'), ('disposition', '"Accept"'),
 ('classifier', '"Not,Spam"'), 
 ('subject', '"=?windows-1255?B?Rlc6IEZ3OiDg5fDp5fog+fno5fog7Pf46eHp7S3u4+Tp7SE=?="')
]

You might want to clean up the quoted strings afterwards (using mystring.strip("'\"")).

EDIT: This regex now also correctly handles escaped quotes inside quoted strings (a="She said \"Hi!\"").

Explanation of the regex:

(\w+)=((?:"(?:\\.|[^\\"])*"|'(?:\\.|[^\\'])*'|[^\\,"'])+)

(\w+): Match the identifier and capture it into backreference no. 1

=: Match a =

(: Capture the following into backreference no. 2:

(?:: One of the following:

"(?:\\.|[^\\"])*": A double quote, followed by either zero or more of the following: an escaped character or a non-quote/non-backslash character, followed by another double quote

|: or

'(?:\\.|[^\\'])*': See above, just for single quotes.

|: or

[^\\,"']: one character that is neither a backslash, a comma, nor a quote.

)+: repeat at least once, as many times as possible.

): end of capturing group no. 2.