BTW is there any way to multiplex/switch the current loop to save board space?.I find Analog Mux useless for the purpose is there any other way ?

I wouldn't do that, either, because any added impedance, for instance by analog switch, would erode symmetry and heavily worsen CMRR.
May be that some relais switching would be possible. But think about heavily increased supply current and board area. No, I think it's not a good idea. But may be someone other has found out a good trick?

Suppose we keep the shield floating at one end now this can also lead to open-circuit voltages which can capacitively couple into our data conductors ?

Exact! But having a shield is always of benefit, even if one end is floating. If CMN contains much high frequent noise, it's wise to parallel R3 and R4 with 100pF NP0 ceramics. You can even increase this capacitances, but keep in mind, that any missmatch will decrease CMRR at higher frequencies. So, when using rather big capacitances only very stable capacitors may be choosen!

Kai