btw - I think I figured that - at least in my example, the range in column E actually seems irrelevant for generating the desired output.
I think also in your formula, Fluff13, appending the range &E4:E8 doesn't really do anything, does it?
So, in the example the following slightly shorter version yields the same result:
And this is also equivalent to:
So, at least my take-away is that the maniupulation of the range passed to SCAN is the important bit rather than to fiddle around too much within the LAMBDA piece which may mess up the accumulator variable.
And then this made me realise that maybe the concatenation and TEXT-xxx part isn't really needed at all:
Bookmarks