diff --git a/doc/release/upcoming_changes/21623.new_feature.rst b/doc/release/upcoming_changes/21623.new_feature.rst new file mode 100644 index 000000000000..fafb2e8d6907 --- /dev/null +++ b/doc/release/upcoming_changes/21623.new_feature.rst @@ -0,0 +1,6 @@ +New parameter ``equal_nans`` added to `np.unique` +----------------------------------------------------------------------------------- + +`np.unique` was changed in 1.21 to treat all ``NaN`` values as equal and return +a single ``NaN``. Setting ``equal_nans=False`` will restore pre-1.21 behavior +to treat ``NaNs`` as unique. Defaults to ``True``. diff --git a/numpy/lib/arraysetops.py b/numpy/lib/arraysetops.py index d44e1a983ebf..6d36fdcbddc9 100644 --- a/numpy/lib/arraysetops.py +++ b/numpy/lib/arraysetops.py @@ -131,13 +131,13 @@ def _unpack_tuple(x): def _unique_dispatcher(ar, return_index=None, return_inverse=None, - return_counts=None, axis=None): + return_counts=None, axis=None, *, equal_nans=None): return (ar,) @array_function_dispatch(_unique_dispatcher) def unique(ar, return_index=False, return_inverse=False, - return_counts=False, axis=None): + return_counts=False, axis=None, *, equal_nans=True): """ Find the unique elements of an array. @@ -162,8 +162,10 @@ def unique(ar, return_index=False, return_inverse=False, return_counts : bool, optional If True, also return the number of times each unique item appears in `ar`. + equal_nans : bool, optional + If True, collapses multiple NaN values in return array into 1 - .. versionadded:: 1.9.0 + .. versionchanged: 1.24 axis : int or None, optional The axis to operate on. If None, `ar` will be flattened. If an integer, @@ -269,7 +271,8 @@ def unique(ar, return_index=False, return_inverse=False, """ ar = np.asanyarray(ar) if axis is None: - ret = _unique1d(ar, return_index, return_inverse, return_counts) + ret = _unique1d(ar, return_index, return_inverse, return_counts, + equal_nans = equal_nans) return _unpack_tuple(ret) # axis was specified and not None @@ -312,13 +315,13 @@ def reshape_uniq(uniq): return uniq output = _unique1d(consolidated, return_index, - return_inverse, return_counts) + return_inverse, return_counts, equal_nans = equal_nans) output = (reshape_uniq(output[0]),) + output[1:] return _unpack_tuple(output) def _unique1d(ar, return_index=False, return_inverse=False, - return_counts=False): + return_counts=False, *, equal_nans=True): """ Find the unique elements of an array, ignoring shape. """ @@ -334,7 +337,8 @@ def _unique1d(ar, return_index=False, return_inverse=False, aux = ar mask = np.empty(aux.shape, dtype=np.bool_) mask[:1] = True - if aux.shape[0] > 0 and aux.dtype.kind in "cfmM" and np.isnan(aux[-1]): + if (equal_nans and aux.shape[0] > 0 and aux.dtype.kind in "cfmM" and + np.isnan(aux[-1])): if aux.dtype.kind == "c": # for complex all NaNs are considered equivalent aux_firstnan = np.searchsorted(np.isnan(aux), True, side='left') else: diff --git a/numpy/lib/tests/test_arraysetops.py b/numpy/lib/tests/test_arraysetops.py index 13385cd2409d..f97fea310dbb 100644 --- a/numpy/lib/tests/test_arraysetops.py +++ b/numpy/lib/tests/test_arraysetops.py @@ -765,3 +765,11 @@ def _run_axis_tests(self, dtype): assert_array_equal(uniq[:, inv], data) msg = "Unique's return_counts=True failed with axis=1" assert_array_equal(cnt, np.array([2, 1, 1]), msg) + + def test_unique_nanequals(self): + # issue 20326 + a = np.array([1, 1, np.nan, np.nan, np.nan]) + unq = np.unique(a) + not_unq = np.unique(a, equal_nans = False) + assert_array_equal(unq, np.array([1, np.nan])) + assert_array_equal(not_unq, np.array([1, np.nan, np.nan, np.nan]))