[issue5669] Extra heapq nlargest/nsmallest option for including ties

[Raymond Hettinger]

Raymond Hettinger <rhettinger at users.sourceforge.net> added the comment:

FWIW, an easy way to pare down millions of entries to dozens is to use a
bigger heap:

from heapq import nsmallest

def nsmallest_with_ties(n, iterable, scale=2):
    ext = n * scale
    s = nsmallest(ext, iterable)
    lastplace = s[n-1]
    if s[-1] == lastplace:
        raise ValueError('may not have found all ties')
    for i in range(n, ext):
        if s[i] != lastplace:
            break
    return s[:i]
    

n = 3
s = [4,3,5,7,4,7,4,3]
print nsmallest_with_ties(3, s)

----------

_______________________________________
Python tracker <report at bugs.python.org>
<http://bugs.python.org/issue5669>
_______________________________________