Possible Duplicate:
python random string generation with upper case letters and digits
How do I generate a String of length X a-z in Python?
Add a comment
|
2 Answers
''.join(random.choice(string.lowercase) for x in range(X))
answered Dec 24, 2009 at 7:54
Ignacio Vazquez-Abrams
804k160 gold badges1.4k silver badges1.4k bronze badges
4 Comments
Zach B
Nice use of generator expressions. But while we're saving memory, might as well use xrange instead of range.
Evan Fosmark
CytokineStorm, as of Python 3.x,
range() behaves the same as xrange().
Federico
Just keep in mind that any sequence generated by the random module is not cryptographically secure.
ShadowRanger
@Federico: Unless you use
random.SystemRandom(), which should be as cryptographically secure as the OS supports. In modules that need cryptographic randomness, I often do: import random, random = random.SystemRandom() to replace the module with an instance of the class that provides OS crypto-randomness. SystemRandom is the random.Random API, with randomness provided by os.urandom instead of Mersenne Twister (os.urandom is backed by stuff like Window's CryptGenRandom, UNIX-like /dev/urandom, etc.).If you want no repetitions:
import string, random
''.join(random.sample(string.ascii_lowercase, X))
If you DO want (potential) repetitions:
import string, random
''.join(random.choice(string.ascii_lowercase) for _ in xrange(X)))
That's assuming that by a-z you mean "ASCII lowercase characters", otherwise your alphabet might be expressed differently in these expression (e.g., string.lowercase for "locale dependent lowercase letters" that may include accented or otherwise decorated lowercase letters depending on your current locale).
3 Comments
u0b34a0f6ae
FYI,
string.lowercase is rarely more than ASCII however you set the locale in my experience. I'm just noting that there is no true variable for "current alphabet".
Anatoly Scherbakov
Neither of these code samples implies uniqueness. Here is a test: pastie.org/8619409
ShadowRanger
@Altaisoft: You're misunderstanding how
sample works. It only draws any given character once, but it's not drawing in order. So abc and cba would be different outputs. Your test is only valid (would have expected huge numbers of overlaps) if you assume sample returns them in the same order they appeared in the original string, but since they can appear in any order, there are (26+7)! / 7! options; overlaps won't be common.