diff options
| author | Julian Taylor <jtaylor.debian@googlemail.com> | 2014-01-19 20:41:42 +0100 |
|---|---|---|
| committer | Julian Taylor <jtaylor.debian@googlemail.com> | 2014-01-23 00:29:55 +0100 |
| commit | cc81da650820e6a623dce41cc66d4b926c7f6941 (patch) | |
| tree | 901d621607aa52a32e12ca225b099768235aa9ab /numpy/polynomial/polyutils.py | |
| parent | e95b6acf10ba8bf4a8c2b7c1841cb1c5f7d6818d (diff) | |
| download | numpy-cc81da650820e6a623dce41cc66d4b926c7f6941.tar.gz | |
BUG: min_scalar_type_num: special float values can be represented in half
E.g. pandas sometime uses float('NaN') which would otherwise convert to
float64 where not required. The new np.where which uses
PyArray_ResultType is a case where the conversion is unnecessary.
Diffstat (limited to 'numpy/polynomial/polyutils.py')
0 files changed, 0 insertions, 0 deletions