Taking advantage of str.split’s behavior with no sep parameter:
>>> s = " \t foo \n bar " >>> "".join(s.split()) 'foobar'
If you just want to remove spaces instead of all whitespace:
>>> s.replace(" ", "") '\tfoo\nbar'
Premature optimization
Even though efficiency isn’t the primary goal—writing clear code is—here are some initial timings:
$ python -m timeit '"".join(" \t foo \n bar ".split())' 1000000 loops, best of 3: 1.38 usec per loop $ python -m timeit -s 'import re' 're.sub(r"\s+", "", " \t foo \n bar ")' 100000 loops, best of 3: 15.6 usec per loop
Note the regex is cached, so it’s not as slow as you’d imagine. Compiling it beforehand helps some, but would only matter in practice if you call this many times:
$ python -m timeit -s 'import re; e = re.compile(r"\s+")' 'e.sub("", " \t foo \n bar ")' 100000 loops, best of 3: 7.76 usec per loop
Even though re.sub is 11.3x slower, remember your bottlenecks are assuredly elsewhere. Most programs would not notice the difference between any of these 3 choices.