Multi-Band Spectroscopy as a Check for RV Signal Origin

There have been a couple planet candidates from Doppler Spectroscopy disproven recently because of the signal that was detected was not consistent across the spectrum. These include a second super-Earth at K2-18 reported by Cloutier, et al. (2017) and shown to be spurious by Sarkis, et al. (2018), and the more recent demonstration by Carleo, et al. (2018) that the hot Jupiter at BD+20 1790 does not exist (though this was suspected as far back as 2010[/url]).

I know the HARPS data downloaded through the ESO server shown to us by Shellface just produces single values for the radial velocity of the star at a given observation, but would it be possible to get wavelength-dependent RVs from the data? I'm guessing we would have to run some analysis of the raw spectrum?

The official HARPS Data Reduction Service can only calculate one radial velocity for a whole spectrum. Some papers using HARPS data do derive wavelength-specific RVs, and I know HARPS-TERRA (which is not publicly available) is capable of calculating them, but I am not sure what the process is. I'd assume it would involve analysing the raw spectrum, as you said, and that would definitely be difficult.