Seems like a bug to me. Certain branches in _array_richcompare return
False to fail rather than Py_NotImplemented, which means the
string-understanding comparison fallbacks don't run. Attached is a
(simple) patch that resolves this bug, and doesn't seem to cause any of
the unit tests to fail. Does this make sense to someone with a better
understanding of the rich comparison code than I?
Mike
On 05/25/2010 12:54 PM, Keith Goodman wrote:
a1 = np.array(['a', 'b'], dtype=object)
>> a2 = np.array(['a', 'b'])
>>
>> a1 == a2
array([ True, True], dtype=bool) # Looks good
>> a2 == a1
False # Should I have expected this?
--
Michael Droettboom
Science Software Branch
Space Telescope Science Institute
Baltimore, Maryland, USA
Index: numpy/core/src/multiarray/arrayobject.c
===================================================================
--- numpy/core/src/multiarray/arrayobject.c (revision 8430)
+++ numpy/core/src/multiarray/arrayobject.c (working copy)
@@ -958,8 +958,8 @@
Py_DECREF(array_other);
if (result == NULL) {
PyErr_Clear();
- Py_INCREF(Py_False);
- return Py_False;
+ Py_INCREF(Py_NotImplemented);
+ return Py_NotImplemented;
}
break;
case Py_NE:
@@ -1021,8 +1021,8 @@
Py_DECREF(array_other);
if (result == NULL) {
PyErr_Clear();
- Py_INCREF(Py_True);
- return Py_True;
+ Py_INCREF(Py_NotImplemented);
+ return Py_NotImplemented;
}
break;
case Py_GT:
@@ -1217,7 +1217,7 @@
}
else if ((strides.ptr == NULL) &&
(buffer.len < (offset + (((intp)itemsize)*
- PyArray_MultiplyList(dims.ptr,
+ PyArray_MultiplyList(dims.ptr,
dims.len))))) {
PyErr_SetString(PyExc_TypeError,
"buffer is too small for " \
Index: numpy/core/tests/test_regression.py
===================================================================
--- numpy/core/tests/test_regression.py (revision 8430)
+++ numpy/core/tests/test_regression.py (working copy)
@@ -1302,5 +1302,13 @@
# Ticket #1345: the following should not cause a crash
np.fromstring(asbytes('aa, aa, 1.0'), sep=',')
+ def test_eq_string_and_object_array(self):
+ # From e-mail thread "__eq__ with str and object" (Keith Goodman)
+ a1 = np.array(['a', 'b'], dtype=object)
+ a2 = np.array(['a', 'c'])
+ assert_array_equal(a1 == a2, [True, False])
+ assert_array_equal(a2 == a1, [True, False])
+
+
if __name__ == "__main__":
run_module_suite()
_______________________________________________
NumPy-Discussion mailing list
[email protected]
http://mail.scipy.org/mailman/listinfo/numpy-discussion
