CODEWITHSUNDEEP
pythonregex
Deepak Yadav
Deepak Yadav

Reputation: 440

Split by space preserving string inside curly braces

str = "cmd -opt1 { a b c  d e f g h } -opt2"

I want output like this: 

[ 'cmd', '-opt1', '{ a b c  d e f g h }', '-opt2' ]

Upvotes: 3

Views: 3012

Answers (4)

Casimir et Hippolyte
Casimir et Hippolyte

Reputation: 89567

In this situation, don't try to split, use re.findall:

>>> import re
>>> re.findall(r'{[^}]*}|\S+', 'cmd -opt1 { a b c  d e f g h } -opt2')
['cmd', '-opt1', '{ a b c  d e f g h }', '-opt2']

if you have to deal with nested curly brackets, the re module doesn't suffice, you need to use the "new" regex module that has the recursion feature.

>>> import regex
>>> regex.findall(r'[^{}\s]+|{(?:[^{}]+|(?R))*+}', 'cmd -opt1 { a b {c d} e f} -opt2')
['cmd', '-opt1', '{ a b {c d} e f}', '-opt2']

Where (?R) refers to the whole pattern itself.

or this one (that is better):

regex.findall(r'[^{}\s]+|{[^{}]*+(?:(?R)[^{}]*)*+}', 'cmd -opt1 { a b {c d} e f} -opt2')

Upvotes: 5

xtofl
xtofl

Reputation: 41509

Do take a look at the argparse module, since I assume you are writing code to parse the arguments of your program. Normally these arguments are stored in sys.argv, so you don't even need to care about splitting the command line string. If you insist on using the command line, you may convert your argument string to an argument list with the str.split method.

import argparse

parser = argparse.ArgumentParser(description='whatever cmd does.')
parser.add_argument('--opt1', metavar='N', type=int, nargs='+',
                   help='integers')

options = parser.parse_args()

for n in options.opt1:
   # do something with n

Upvotes: 1

James Mills
James Mills

Reputation: 19040

Just split on the { and } then split the separate parts by a regular space:

str = "cmd -opt1 { a b c d e f g h } -opt2"
>>> a, b = str.split("{")
>>> c, d = b.split("}")
>>> a.split() + ["{{{0}}}".format(c)] + d.split()
['cmd', '-opt1', '{ a b c d e f g h }', '-opt2']

Upvotes: 0

vks
vks

Reputation: 67968

\s+(?![^{]*})

You can split by this.See demo.

https://regex101.com/r/jV9oV2/6

Upvotes: 3

Related Questions