RegEx for matching multiple substrings using one group?

Question

I'm trying to parse a string with unknown number of elements with RegEx in Python. Here is the example:

>>>> import re
>>>> re.match("\=( A([0-9]+))*", "= A1 A2 A3 A4").groups()[1::2]
('4',)

I expect to have:

('1', '2', '3', '4',)

How can I get the expected result?

EDIT:

re.findall will not work for me. Let me make a better example:

I want to match the following string:

_func(cmd, param1, param2, param3, param4)_

I don't know in advance the nummber of parameters. I expected to solve it using the following code:

>>> re.match("(\w+)\(cmd(, (\w+))*\)", "func(cmd, param1, param2, param3, param4)")

But this do not work, since groups ()* are not expanded to many items, but only last is used. Any ideas?

install ipython (pip install ipython), import 're' module (import re), and look at the docs (?re.match, ?re.search, ?re.findall). — Vidul
– Vidul, Commented Jun 4, 2012 at 23:10
Your comment was quite useless. Do you think I've not read all these before asking? — PoltoS
– PoltoS, Commented Jun 4, 2012 at 23:22
-1. yes, I think so, your question leads to this conclusion. I'd recommend you to leave your ignorance at home and absorb more polite attitude. — Vidul
– Vidul, Commented Jun 4, 2012 at 23:58
@Vidul Petrov: There are some rules on Stack Overflow. If you know the answer - answer, if you don't - don't, if you don't understand the question - ask for more info. Don't point users to the manual without any answer (it's unpolite!). After all my question is not trivial and it was found that there is no direct solution. And you have not presented an answer at all! And finaly, on StackOverflow we do not -1 questions just because we are unable to answer and we point users to the manual without any solution. — PoltoS
– PoltoS, Commented Jun 5, 2012 at 14:06

steveha · Accepted Answer · 2012-06-04 23:25:23Z

3

pat = re.compile(r' A(\d+)')
lst = re.findall(pat, "= A1 A2 A3 A4")

This returns a list, and in your example you showed a tuple. I presume a list will work for you, but of course you can always do:

t = tuple(lst)

The answer I just gave doesn't actually check for the = in the input string. If you need to do that, you can always use two patterns and two steps:

pat0 = re.compile(r'=(?: A\d+)+')
pat1 = re.compile(r' A(\d+)')

m = pat0.search("= A1 A2 A3 A4")
if not m:
    print("input string not what was expected")
else:
    s = m.group(0)
    lst = re.findall(pat, s)

EDIT: Code that handles your func() example:

s_code = "func(cmd, param1, param2, param3, param4)"
pat_recognize_args = re.compile(r'func\(cmd([^)]*)\)')
pat_parse_args = re.compile(r'[, ]+([^, ]+)')

m = pat_recognize_args.search(s_code)
if m:
    s = m.group(1)
    lst = re.findall(pat_parse_args, s)

When I ran the above code, lst was set to: ['param1', 'param2', 'param3', 'param4']

pat_recognize_args looks for the literal string func with a literal ( (which is backslash-escaped in the pattern so re won't try to use it to start a match group), then the literal string cmd, and then a match group that matches anything up to a literal ) character; then the match group is closed with a ) and a literal ) is there to match the actual ) that finishes the function call. After this pattern matches, the match object will have group 1 set to just the interesting arguments from the function call.

So next we set s = m.group(1) and then have re.findall() pull out the arguments for us.

edited Jun 4, 2012 at 23:25

answered Jun 4, 2012 at 23:06

steveha

77.1k21 gold badges94 silver badges119 bronze badges

Sign up to request clarification or add additional context in comments.

6 Comments

PoltoS Over a year ago

This will not work for me, since I need to match the = sign too. I'll edit my question to be more presize.

PoltoS Over a year ago

Isn't there a way to do it in one? My original RegEx is much more complex to split it in several matches: re.match("// JS/Py API prototype: *(\w+)\(self( *, *(\w+)( *\= *([0-9]+|None|False|True))?)*\)", functionPrototype)

PoltoS Over a year ago

After googling more I've found that it is not possible to do in one step in Python. Perl 6 and .NET provides a way to do it, but that's a specific extention of RegEx, which is not standartize. Thanks for your two-steps answer

steveha Over a year ago

It is easy to make a match group that collects a whole bunch of stuff into one string. It is easy to make a match group that collects one thing. But I do not know any way to make a pattern that returns a variable number of match groups. I'm not sure why you want to do it in one pattern, because it seems like the painful stuff is just in recognizing the function call, while collecting the args is pretty simple (just split on commas, basically). And regular expressions can be complicated enough without trying to be as tricky as possible; better to use two simpler patterns than one monster.

xvatar Over a year ago

@steveha nice answer, i have one question: why we need the ?: so that it works?

|

Collectives™ on Stack Overflow

RegEx for matching multiple substrings using one group?

1 Answer 1

6 Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

1 Answer 1

6 Comments

Your Answer

Sign up or log in

Post as a guest

Related