This puzzles me,

> preserve all alphabetic elements of the array but the duplicate numeric elements must be removed

but I think what you want is to parse the data pairwise ($number, $name) and have a unique list of names per number.

In this case I'd suggest building a hash of hashes (if original order doesn't matter). Just set

$names_per_num{$number}{$name} = 1

for each combination.

After that you'll just need to iterate over all numbers and print the keys of the sub-hash to get your desired output.

No code yet, we'd love to help you improving your attempts! :)

HTH!

PS: and if original order matters just use the above HoH as a %seen filter while iterating the list.