Python re: Storing multiple matches in variables

Question

I want to match different parts of a string and store them in separate variables for later use. For example,

string = "bunch(oranges, bananas, apples)"
rxp = "[a-z]*\([var1]\, [var2]\, [var3]\)"

so that I have

var1 = "oranges"
var2 = "bananas"
var3 = "apples"

Something like what re.search() does but for multiple different parts of the same match.

EDIT: the number of fruits in the list is not known beforehand. Should have put this in with the question.

Answer 1

That is what re.search does. Just use capturing groups (parentheses) to access the stuff that was matched by certain subpatterns later on:

>>> import re
>>> m = re.search(r"[a-z]*\(([a-z]*), ([a-z]*), ([a-z]*)\)", string)
>>> m.group(0)
'bunch(oranges, bananas, apples)'
>>> m.group(1)
'oranges'
>>> m.group(2)
'bananas'
>>> m.group(3)
'apples'

Also note, that I used a raw string to avoid the double backslashes.

If your number of "variables" inside bunch can vary, you have a problem. Most regex engines cannot capture a variable number of strings. However in that case you could get away with this:

>>> m = re.search(r"[a-z]*\(([a-z, ]*)\)", string)
>>> m.group(1)
'oranges, bananas, apples'
>>> m.group(1).split(', ')
['oranges', 'bananas', 'apples']

Answer 2

For regular expressions, you can use the match() function to do what you want, and use groups to get your results. Also, don't assign to the word string, as that is a built-in function (even though it's deprecated). For your example, if you know there are always the same number of fruits each time, it looks like this:

import re
input = "bunch(oranges, bananas, apples)"
var1, var2, var3 = re.match('bunch\((\w+), (\w+), (\w+)\)', input).group(1, 2, 3)

Here, I used the \w special sequence, which matches any alphanumeric character or underscore, as explained in the documentation

If you don't know the number of fruits in advance, you can use two regular expression calls, one to get extract the minimal part of the string where the fruits are listed, getting rid of "bunch" and the parentheses, then finditer to extract the names of the fruits:

import re
input = "bunch(oranges, bananas, apples)"
[m.group(0) for m in re.finditer('\w+(, )?', re.match('bunch\(([^)]*)\)', input).group(1))] 

Answer 3

If you want, you can use groupdict to store matching items in a dictionary:

regex = re.compile("[a-z]*\((?P<var1>.*)\, (?P<var2>.*)\, (?P<var3>.*)")
match = regex.match("bunch(oranges, bananas, apples)")
if match:
    match.groupdict()

#{'var1': 'oranges', 'var2': 'bananas', 'var3': 'apples)'}

Answer 4

Don't. Every time you use var1, var2 etc, you actually want a list. Unfortunately, this is no way to collect arbitrary number of subgroups in a list using findall, but you can use a hack like this:

import re
lst = []
re.sub(r'([a-z]+)(?=[^()]*\))', lambda m: lst.append(m.group(1)), string)
print lst # ['oranges', 'bananas', 'apples']

Note that this works not only for this specific example, but also for any number of substrings.