Ah there are two conditions that the current graph will fail by the List.Deconstruct node:
- Sheets with no revisions, because they produce an Empty List
- Sheets with only one revision, because they produce a list with only one item
The graph I shared above failed to account for these conditions because, well, my sample file didn’t have them, oops! 
I think the best way to handle this would be to separate sheets with no revisions, sheets with only one revision, and sheets with multiple revisions, then join their data back together before writing to Excel. See the below graph (the non-green colored groups are the new additions since the last graph, and I colored the groups by the conditions they’re handling)
Additionally - the original graph only appeared to output the sheets sorted by their sheet number, because our sample files more than likely had each sheet element created in the same order as their sheet number. In actuality, the graph was by default sorted by element ID which follows order of creation. I added a bit towards the end just before flattening to write to Excel that uses Sort By Key to sort the lists by sheet number. I definitely suggest going through and doing all of the exercises in the Dynamo Primer once you’ve finished this graph 