This idea had to be implemented 15 yeas ago, but we are still waiting. LabVIEW has many such small handicaps with arrays obstructing efficient programming: another example is low fuctionality of array nodes for in-place-element structure.
For completeness and showing the different kinds of possible implementations, I am linking to a page on LAVA discussing the issue: Array auto-indexing It is so far the oldest public writing about the suggested idea. (Thank you GregSands!) The picture is so beautiful, that I have to re-link it, too: