While working on #929, I realised it would be very handy to have a way to change the unit of a quantity into a compatible one, and recalculate the values it has in-place. There are probably multiple ways to do this (@wkerzendorf suggested to(...,inplace=True), following PANDAS; one could also think of .to(..., out=one-self), following numpy ufuncs), but for an initial trial, I simply allowed assignment to .unit, which will convert to the requested unit if possible:

    def get_unit(self):
        return self._unit

    def set_unit(self, new_unit):
        self_values = self.__array__()
        self_values *= self.unit.to(new_unit)
        self._unit = Unit(new_unit)

    unit = property(get_unit, set_unit)

(Here, I have to get the array view to ensure not to get the overloaded multiply.)
With this,

import astropy.units as u; import numpy as np
a = np.arange(0.,100.,10)*u.deg
a.unit = u.rad
print(a)

yields

[ 0.          0.02777778  0.05555556  0.08333333  0.11111111 
0.13888889 0.16666667  0.19444444  0.22222222  0.25      ] cycle

As indicated above, I'm not wedded to a particular implementation (though I'm happy to turn this into a PR), but I think we do need the functionality (or is it there already? @iguananaut, @astrofrog?)