Python-分箱

Question

I'm generating a series of values and would like to bin them. I'd rather not use numpy or the like. Is there something more pythonic than:

bins = [20,30,40]
results = [0,0,0,0]

for _ in range(iterations):
    x = somefunction()
    for n, bin in enumerate(bins):
        if x < bin:
            results[n] += 1
            break
    else:
        results[-1] += 1

final = [100 * r / float(iterations) for r in results]
print(final)

Answer 1

You could use something like that:

r = [0] * (len(bins) + 1)
for _ in xrange(iterations):
    r[next((i for i, bin in enumerate(bins) if somefunction() < bin), -1)] += 1

or alternatively a counter:

n = len(bins)
from collections import Counter
c = Counter()
c.update(
    next((i for i, bin in enumerate(bins) if somefunction() < bin), n)
    for _ in xrange(iterations)
)

Answer 2

It would be better/faster (for larger arrays) to use a binary search algorithm instead of a linear search algorithm.

That is,

def binm(rr,ra):
  ih=len(ra)-1
  il=0
  if rr<ra[il]: return il
  while (ih-il>1):
    ie=(ih+il)/2
    if rr<ra[ie]:
      ih=ie
    else:
      il=ie
  return ih

bins = [20,30,40]
results = [0,0,0,0]

for _ in range(iterations):
  x = somefunction()
  ib=binm(x,bins)
  results[ib]+=1