diff options
| author | mattip <matti.picus@gmail.com> | 2019-10-13 13:50:23 +0300 |
|---|---|---|
| committer | mattip <matti.picus@gmail.com> | 2019-10-13 13:50:23 +0300 |
| commit | e6e0ac11345c3637e7b012ee8ab10e70461525bc (patch) | |
| tree | 9fdcf1cd472ac859f1f0f411eec2ed0b176644a1 /doc/neps | |
| parent | 6dc23b070cc66d10e6c03946483b33b176ecb6d8 (diff) | |
| download | numpy-e6e0ac11345c3637e7b012ee8ab10e70461525bc.tar.gz | |
NEP: fixes from review, add ragged_array_object function
Diffstat (limited to 'doc/neps')
| -rw-r--r-- | doc/neps/nep-0034.rst | 32 |
1 files changed, 22 insertions, 10 deletions
diff --git a/doc/neps/nep-0034.rst b/doc/neps/nep-0034.rst index b7b9d048e..91ee2c0ee 100644 --- a/doc/neps/nep-0034.rst +++ b/doc/neps/nep-0034.rst @@ -1,6 +1,6 @@ -========================================== -NEP 34 — Disallow default ``dtype=object`` -========================================== +=========================================================== +NEP 34 — Disallow inferring ``dtype=object`` from sequences +=========================================================== :Author: Matti Picus :Status: Draft @@ -11,8 +11,8 @@ NEP 34 — Disallow default ``dtype=object`` Abstract -------- -``np.array([<value])`` with no ``dtype`` keyword argument will sometimes -default to an ``object``-dtype array. Change the behaviour to raise a +``np.array([<sequence>, <sequence>])`` with no ``dtype`` keyword argument will +sometimes default to an ``object``-dtype array. Change the behaviour to raise a `ValueError` instead. Motivation and Scope @@ -31,19 +31,31 @@ Detailed description After this change, ragged array creation must explicitly define a dtype: - a = np.array([1, 2], [1]) + np.array([[1, 2], [1]]) ValueError: cannot guess the desired dtype from the input - a = np.array([1, 2], [1], dtype=object) - print(a.dtype) - object + np.array([[1, 2], [1]], dtype=object) + ValueError: cannot guess the desired object type, use \ + ``ragged_object_array`` instead + + np.ragged_array_object([[1, 2], [1]], depth=0) + array([list([[1, 2], [1]])], dtype=object) + + np.ragged_array_object([[1, 2], [1]], depth=1) + array([list([1, 2]), list([1])], dtype=object) + + np.ragged_array_object([[1, 2], [1]], depth=2) + ValueError: cannot interpret input as depth-2 sequences + + np.ragged_array_object([[[1, 2], [1]], [['a']]], depth = 2) + array([[list([1, 2]), list([1])], [list(['a'])]], dtype=object) Related Work ------------ `PR 14341`_ tried to raise an error when ragged arrays were specified with a numeric dtype ``np.array, [[1], [2, 3]], dtype=int)`` but failed due to -false-postives, for instance ``np.array([1, np.array([5])], dtype=int)``. +false-positives, for instance ``np.array([1, np.array([5])], dtype=int)``. .. _`PR 14341`: https://github.com/numpy/numpy/pull/14341 |