When I call:
np.fromstring('3 3 3 0', sep=' ')
it returns
array([ 3., 3., 3., 0.])
Since by default, sep='', I would expect the following call to return the same result:
np.fromstring('3330')
However it raises
ValueError: string size must be a multiple of element size
Why is this? What's the most pythonic way of getting array([ 3., 3., 3., 0.]) from '3330'?
You could use np.fromiter:
In [11]: np.fromiter('3300', dtype=np.float64)
Out[11]: array([ 3., 3., 0., 0.])
In [12]: np.fromiter('3300', dtype=np.float64, count=4) # count=len(..)
Out[12]: array([ 3., 3., 0., 0.])
list('3330') turns the string into a 4 element list of chars; with the correct dtype, np.array will convert those strings to numbers:
In [48]: np.array(list('3330'),dtype=float)
Out[48]: array([ 3., 3., 3., 0.])
The fromiter solution is faster, but that's one more function to remember.
sepkeyword: "If not provided or, equivalently, the empty string, the data will be interpreted as binary data."