diff options
| author | Mark Wiebe <mwiebe@enthought.com> | 2011-08-02 17:06:20 -0500 |
|---|---|---|
| committer | Charles Harris <charlesr.harris@gmail.com> | 2011-08-27 07:26:50 -0600 |
| commit | 00dff152bb10c810176c7e47ad41e563df602208 (patch) | |
| tree | 156498d06b040209f4581edb1b91f620c5b98afe /numpy/lib/tests | |
| parent | 0375181e807f9def0e1bd0279214f5a00ec55485 (diff) | |
| download | numpy-00dff152bb10c810176c7e47ad41e563df602208.tar.gz | |
ENH: umath: Make sum, prod, any, or functions use the .reduce method directly
This required some extensive changes, like:
* Making logical_or, logical_and, and logical_not on object arrays behave
like their Python equivalents instead of calling methods on the objects
* Changing the units for a fair number of the binary operations to
None, so they don't get initialized to their unit values at the start
A consequence of this is that multi-dimensional reductions like sum, prod,
any, or all no longer need to make copies, so are faster in some cases.
Diffstat (limited to 'numpy/lib/tests')
| -rw-r--r-- | numpy/lib/tests/test_function_base.py | 2 |
1 files changed, 1 insertions, 1 deletions
diff --git a/numpy/lib/tests/test_function_base.py b/numpy/lib/tests/test_function_base.py index 105389d6d..0b6b1e19d 100644 --- a/numpy/lib/tests/test_function_base.py +++ b/numpy/lib/tests/test_function_base.py @@ -452,7 +452,7 @@ class TestTrapz(TestCase): def test_simple(self): r = trapz(exp(-1.0 / 2 * (arange(-10, 10, .1)) ** 2) / sqrt(2 * pi), dx=0.1) #check integral of normal equals 1 - assert_almost_equal(sum(r, axis=0), 1, 7) + assert_almost_equal(r, 1, 7) def test_ndim(self): x = linspace(0, 1, 3) |