in reply to
Re^2: how to merge many files of sorted hashes?
in thread how to merge many files of sorted hashes?
Having seen your code here it's now clear that a given key can show up in multiple files and therefore you probably do want to be merging those values and only having the key appear in one place, which then means that a certain amount of copying is unavoidable.
There's also an advantage to having output files that hold every key within a certain range, which then means your master file/table doesn't need to have a column saying which file a key is in (i.e., since you just infer it from the range the key is in), which then saves massive amounts of space...