[Numpy-svn] r3317 - trunk/numpy/lib
numpy-svn at scipy.org
numpy-svn at scipy.org
Thu Oct 12 07:19:04 EDT 2006
Author: rc
Date: 2006-10-12 06:19:01 -0500 (Thu, 12 Oct 2006)
New Revision: 3317
Modified:
trunk/numpy/lib/arraysetops.py
Log:
Modified: trunk/numpy/lib/arraysetops.py
===================================================================
--- trunk/numpy/lib/arraysetops.py 2006-10-12 11:14:55 UTC (rev 3316)
+++ trunk/numpy/lib/arraysetops.py 2006-10-12 11:19:01 UTC (rev 3317)
@@ -11,17 +11,17 @@
union1d,
setdiff1d
-All functions work best with integer numerical arrays on input. For floating
-point arrays, innacurate results may appear due to usual round-off and
-floating point comparison issues.
+All functions work best with integer numerical arrays on input
+(e.g. indices). For floating point arrays, innacurate results may appear due to
+usual round-off and floating point comparison issues.
Except unique1d, union1d and intersect1d_nu, all functions expect inputs with
unique elements. Speed could be gained in some operations by an implementaion
of sort(), that can provide directly the permutation vectors, avoiding thus
calls to argsort().
-Run test_unique1d_speed() to compare numpy.unique1d() with numpy.unique() - it
-should be the same.
+Run test_unique1d_speed() to compare performance of numpy.unique1d() and
+numpy.unique() - it should be the same.
To do: Optionally return indices analogously to unique1d for all functions.
@@ -33,7 +33,6 @@
__all__ = ['ediff1d', 'unique1d', 'intersect1d', 'intersect1d_nu', 'setxor1d',
'setmember1d', 'union1d', 'setdiff1d']
-
import time
import numpy as nm
@@ -202,7 +201,6 @@
assert nm.alltrue( b == c )
-
print nItems
print dt1s
print dt2s
More information about the Numpy-svn
mailing list