We strip (whitespace from) labels in some cases but not in others
In a axis creation string, they are stripped (seems fine to me)
>>> arr = ndtest('a=a0,a1 ,a2')
>>> arr.axes
AxisCollection([
Axis(['a0', 'a1', 'a2'], 'a')
])
In explicit axis creation, they are not stripped (still fine to me)
>>> arr = ndtest(Axis(['a0', 'a1 '], 'a'))
>>> arr.axes
AxisCollection([
Axis(['a0', 'a1 '], 'a')
])
When targeting an explicit key on an explicit axis, they are stripped. This is probably a bad idea.
>>> arr[arr.a['a1 ']]
ValueError: a['a1'] is not a valid label for any axis
>>> arr.a['a1 ']
a['a1']
>>> arr[arr.a.i[1]]
1
>>> arr.a[['a1 ', 'a2']]
a['a1 ', 'a2']
We should either strip them during axis creation (with a warning) OR not strip them when targeting a single label on an axis.
We could implement the following:
>>> arr.a['a1 ']
a['a1 ']
>>> arr.a['a1 ,a2']
a['a1', 'a2']
>>> arr.a[['a1 ', 'a2']]
a['a1 ', 'a2']
We strip (whitespace from) labels in some cases but not in others
In a axis creation string, they are stripped (seems fine to me)
In explicit axis creation, they are not stripped (still fine to me)
When targeting an explicit key on an explicit axis, they are stripped. This is probably a bad idea.
We should either strip them during axis creation (with a warning) OR not strip them when targeting a single label on an axis.
We could implement the following: