It depends on whether the light sabre is "optically thick" to light. For a fully ionised plasma and a continuum light source, you would be relying on Thomson scattering from free electrons to provide opacity. The cross-section per electron is $\sigma = 6.6\times 10^{-29}$ m$^2$ and independent of wavelength.
The mean free path of light in the plasma is $(n\sigma)^{-1}$, where $n$ is the electron number density. This then has to be smaller than the width of the sabre "beam" if it is to block the light and cast a shadow. If we imagine the "beam" width is about 5cm, then $ n> (0.05\sigma)^{-1} = 3\times 10^{29}$ m$^{-3}$.
If the plasma is pure ionised hydrogen, then it would have a density of 500 kg/m$^3$, so I guess this is just about conceivable/portable.
Of course you could cheat. For example if you take a partially ionised plasma and then only illuminate it with a lamp that emits photons at the discrete resonant frequency corresponding to a particular atomic/ionic transition in the plasma, then it could be much sparser and cast a shadow because resonant cross-sections are much greater than the Thomson scattering cross-section.