On Sat, Jul 11, 2020 at 1:33 PM David Mertz <mertz@gnosis.cx> wrote:
On Sat, Jul 11, 2020 at 3:45 PM Christopher Barker <pythonchb@gmail.com> wrote:
random.choice(the_dict.keys())
is a little easier than:
random.choice(list(the_dict.keys())
Ummm... don't you mean:
random.choice(list(the_dict))
well, sure, though I have to say that I think that that's an unfortunate confusing thing about python dicts. IN fact, I doubt there are many uses at all for dict.keys() -- most uses can jsut use the dict. I certainly see newbies write: if this in dict.keys(): pretty often. maybe adding indexing to the dict views will give the dict_keys object more reason to exist :-) -CHB
If it's keys you care about I've saved you one character over your proposed style, while also reading better to me. It's only for .items() where it doesn't work. And honestly, just looking up the value from the random key is not hard.
In any case, if "reservoir sampling" is the goal here, we should just add a function `random.reservoir_sample()` to accommodate using iterators rather than sequences (https://en.wikipedia.org/wiki/Reservoir_sampling)
-- The dead increasingly dominate and strangle both the living and the not-yet born. Vampiric capital and undead corporate persons abuse the lives and control the thoughts of homo faber. Ideas, once born, become abortifacients against new conceptions.
-- Christopher Barker, PhD Python Language Consulting - Teaching - Scientific Software Development - Desktop GUI and Web Development - wxPython, numpy, scipy, Cython