New submission from Raymond Hettinger <raymond.hettin...@gmail.com>:
Weighted sampling without replacement isn't easy for a user to do with the
current tooling. There is a StackOverflow question about how to do it. Also,
this service is currently offered by numpy.
Use it like this:
>>> sample(['katniss', 'prim', 'gale', 'peeta'] , weights=[20, 1, 42, 10],
k=2)
['prim', 'peeta']
Here's an efficient implementation similar to how numpy does it:
--- a/Lib/random.py
+++ b/Lib/random.py
@@ -331,7 +331,7 @@ class Random(_random.Random):
j = _int(random() * (i+1))
x[i], x[j] = x[j], x[i]
- def sample(self, population, k, *, counts=None):
+ def sample(self, population, k, *, counts=None, weights=None):
"""Chooses k unique random elements from a population sequence or set.
Returns a new list containing elements from the population while
@@ -392,6 +392,18 @@ class Random(_random.Random):
if not isinstance(population, _Sequence):
raise TypeError("Population must be a sequence. For dicts or
sets, use sorted(d).")
n = len(population)
+ if weights is not None:
+ if counts is not None:
+ raise TypeError('Cannot specify both counts and weights')
+ weights = list(weights)
+ positions = range(n)
+ indices = []
+ while (needed := k - len(indices)):
+ for i in choices(positions, weights, k=needed):
+ if weights[i]:
+ weights[i] = 0.0
+ indices.append(i)
+ return [population[i] for i in indices]
if counts is not None:
cum_counts = list(_accumulate(counts))
if len(cum_counts) != n:
----------
components: Library (Lib)
messages: 368458
nosy: mark.dickinson, rhettinger, tim.peters
priority: normal
severity: normal
status: open
title: Add optional weights to random.sample()
type: enhancement
versions: Python 3.9
_______________________________________
Python tracker <rep...@bugs.python.org>
<https://bugs.python.org/issue40569>
_______________________________________
_______________________________________________
Python-bugs-list mailing list
Unsubscribe:
https://mail.python.org/mailman/options/python-bugs-list/archive%40mail-archive.com
