summaryrefslogtreecommitdiff
path: root/doc
diff options
context:
space:
mode:
authorSebastian Berg <sebastian@sipsolutions.net>2022-11-04 19:23:51 +0100
committerGitHub <noreply@github.com>2022-11-04 11:23:51 -0700
commitd2d6c39d73713cb922cec4b0f37db64d985b83a0 (patch)
tree8cad42f99f2699a0f8aa8de4d8b55610c86ae62e /doc
parent77728885a04a3178909111e7ce3f96a544b2796c (diff)
downloadnumpy-d2d6c39d73713cb922cec4b0f37db64d985b83a0.tar.gz
DOC: Remove "current" from byte-order note and expand it slightly (#22521)
We don't preserve non-native byte-order and we have no wish to do so. (If someone disagrees, they better wite a NEP ;)) Spell it out, and thus: closes gh-1866 Co-authored-by: Sebastian Berg <sebastianb@nvidia.com>
Diffstat (limited to 'doc')
-rw-r--r--doc/source/user/byteswapping.rst8
1 files changed, 5 insertions, 3 deletions
diff --git a/doc/source/user/byteswapping.rst b/doc/source/user/byteswapping.rst
index ddc375683..26ffa8943 100644
--- a/doc/source/user/byteswapping.rst
+++ b/doc/source/user/byteswapping.rst
@@ -61,13 +61,15 @@ is also big-endian). However, sometimes you need to flip these around.
.. warning::
- Scalars currently do not include byte order information, so extracting
- a scalar from an array will return an integer in native byte order.
- Hence:
+ Scalars do not include byte order information, so extracting a scalar from
+ an array will return an integer in native byte order. Hence:
>>> big_end_arr[0].dtype.byteorder == little_end_u4[0].dtype.byteorder
True
+ NumPy intentionally does not attempt to always preserve byte-order
+ and for example converts to native byte-order in `numpy.concatenate`.
+
Changing byte ordering
======================