Repository navigation
BUG: Converting masked array to bytes or unicode gives 0.0 #9750
Description
This is a little worrying
>>> a_u = np.zeros((), 'U10')
>>> a_u[()] = np.ma.masked
>>> a_u[()]
u'0.0'>>> a_b = np.zeros((), 'S10')
>>> a_b[()] = np.ma.masked
>>> a_b[()]
b'0.0'This is a regression introduced by #8903. Prior to that, it seems that unicode arrays produced u'--'
Activity
[-]BUG: Converting masked array to bytes gives `0.0`[/-][+]BUG: Converting masked array to bytes or unicode gives `0.0`[/+]