Terry J. Reedy <tjre...@udel.edu> added the comment:
Luís, what you really need for the problem you outline is an immutable set with
one operation, 'in', aside from create and delete. If specialized to strings
only, it could internally stores only character sequences, rather than Python
objects. I suspect someone has written such a class (in C), but did not find
anything on pypi.python.org.
Lacking that, the following experiment suggests that you might be able to
restore near O(n) time by partitioning the set of keys into a collection of
sets. My 16M set took 6.48 and 1.80 seconds. Times 4 is 25.9 and 7.2. My 64M
set took 28 and 19 seconds, but 4 16M sets take 26.3 and 7.5 seconds, only
about 5% more than the x4 target.
print(timeit.Timer(
's = [{str(n) for n in range(i*1000000, (i+16)*1000000)}'
' for i in (0, 16, 32, 48)]')
.timeit(number=1))
print(timeit.Timer(
'del s',
's = [{str(n) for n in range(i*1000000, (i+16)*1000000)}'
' for i in (0, 16, 32, 48)]')
.timeit(number=1))
----------
_______________________________________
Python tracker <rep...@bugs.python.org>
<https://bugs.python.org/issue32846>
_______________________________________
_______________________________________________
Python-bugs-list mailing list
Unsubscribe:
https://mail.python.org/mailman/options/python-bugs-list/archive%40mail-archive.com
