From: W. Trevor King <wking@drexel.edu>
Date: Fri, 29 Oct 2010 19:19:31 +0000 (-0400)
Subject: Use numpy.ndarray.sum() vs. __builtin__.sum() to compute hist bin counts.
X-Git-Url: http://git.tremily.us/?a=commitdiff_plain;h=8bb96b14de3e4a8da0bbb3c378f84f601e9e87e6;p=sawsim.git

Use numpy.ndarray.sum() vs. __builtin__.sum() to compute hist bin counts.

It's much faster:
  $ time python -c 'import numpy; sum(numpy.arange(int(1e6)))'
  real    0m3.312s
  user    0m3.188s
  sys     0m0.118s
  $ time python -c 'import numpy; numpy.arange(int(1e6)).sum()'
  real    0m0.878s
  user    0m0.753s
  sys     0m0.122s
---

diff --git a/pysawsim/histogram.py b/pysawsim/histogram.py
index 5b4480b..bb66c05 100644
--- a/pysawsim/histogram.py
+++ b/pysawsim/histogram.py
@@ -69,7 +69,7 @@ class Histogram (object):
         bin_is = numpy.floor((data - self.bin_edges[0])/bin_width)
         self.counts = []
         for i in range(len(self.bin_edges)-1):
-            self.counts.append(sum(bin_is == i))
+            self.counts.append(sum(bin_is == i).sum())
         self.total = float(len(data)) # some data might be outside the bins
         self.mean = data.mean()
         self.std_dev = data.std()