Sensitivity curves of negative film are different from the permissivity curves of the filters used on digital camera sensors. For example, film was decidedly sensitive to UV light, making the use of UV filters (and also skylight filters) a good idea for certain film situations. With digital, UV filters are basically a wash and don't really cause additional changes (in theory, they might affect purple fringing but I haven't been able to see a noticeable effect: probably they don't do more than the sensor filters already do).
Now fluorescent lights tend to have very pronounced lines in their visible light spectrum. While for the broad spectrums of incandescent light sources and typical reflective colored surfaces the responses of film, human eye, and digital sensor filters can be pretty well dealt with by simple linear mappings, the narrow lines of fluorescent lights are much more problematic. If the relative sensitivities at those wavelengths are not similar to the relative sensitivities of what the sensors broadly cover, you get a problem.
A similar problem is matching paint swatches: if you perfectly match a car's color in sunshine for a paint repair job, that is no guarantee that the result will not stick out like a sore thumb under sodium vapor or fluorescent lighting. Incandescent, in contrast, will usually look fine.
Good correction filters will be a good compromise for various forms of light, but for line spectra there just is not much room for compromise.