Python split string by start and end characters

Question

Say you have a string like this: "(hello) (yes) (yo diddly)".

You want a list like this: ["hello", "yes", "yo diddly"]

How would you do this with Python?

Solution 1

I would do it like this:

"(hello) (yes) (yo diddly)"[1:-1].split(") (")

First, we cut off the first and last characters (since they should be removed anyway). Next, we split the resulting string using ") (" as the delimiter, giving the desired list.

Solution 2

import re
pattern = re.compile(r'\(([^)]*)\)')

The pattern matches the parentheses in your string (\(...\)) and these need to be escaped.
Then it defines a subgroup ((...)) - these parentheses are part of the regex-syntax.
The subgroup matches all characters except a right parenthesis ([^)]*)

s = "(hello) (yes) (yo diddly)"
pattern.findall(s)

gives

['hello', 'yes', 'yo diddly']

UPDATE:
It is probably better to use [^)]+ instead of [^)]*. The latter would also match an empty string.

Using the non-greedy modifiers, as DSM suggested, makes the pattern probably better to read: pattern = re.compile(r'\((.+?)\)')

Solution 3

This will give you words from any string :

>>> s="(hello) (yes) (yo diddly)"
>>> import re
>>> words = re.findall(r'\((.*?\))',s)
>>> words
['hello', 'yes', 'yo diddly']

as D.S.M said.

? in the regex to make it non-greedy.

Comments

[Name McChange]

I've got to wait 6 minutes, just hold on ;)

[arshajii]

I think he would want "yo diddly" as one entry in the array.

[vivek]

@DSM you dont need the inner pair of parenthesis.

[DSM]

@vivek: but if you don't use them, then you get left with parentheses. Compare your output to the desired output.

[Roland Smith]

The answer using regular expressions is more flexible. This answer would not produce the correct output e.g. if the string started or ended with a space, or if there was more than one space or a tab character between the parentheses.

[DSM]

Does this have advantages over using ??

[arshajii]

Very true. I purposefully avoided regular expressions in this answer because I thought a solution could be achieved through simpler means. Of course, the OP can use them if he requires further flexibility like you mentioned.

[tzelleke]

@DSM I don't know - but probably its clearer to use the non-greedy syntax. I'll update my answer this way