at CADC we chose the band that intersects the largest fraction of the Plane.energy.bounds interval aka "this is mostly UV"... no one ever noted or complained about this
it would be feasible change to Plane.energy.emBand [0..2] so that there could be 0 (unknown), 1, or 2 band names; in the case of 2 it would be lower and upper and not include additional bands in the middle (e..g UV-to-Infrared rather than UV-and-Optical-and-Infrared)... serialised in the database could use a single column and | separator (like keywords)
So: feasible = yes, useful = ???
allowing a List value would not invalidate existing metaChecksums since the checksum of a single value is the same if it is in a list (a null and an empty list both contribute nothing and a value and a list of 1 value contribute the same bytes to the metaChecksum of the Plane)
Some filters span multiple predefined energy regimes (eg HST filters span optical and UV).