Thanks for your suggestion! For what use case do you need this?

This is nice to check if the column retrieved in xlwings actually matches the intended column that exists in the associated open Microsoft Excel program.

Another use case is to specify a range based on another computed range:

col_letter = ws["A1"].expand().last_cell.offset(row_offset=0, column_offset=1).address.split("$")[1]
row_number = 5

ws.range("A2:" + str(col_letter) + str(row_number))

I also just realized that the following is the same:
sheet["A1"].address.split("$")[1] == sheet["A1"].address[1]

sheet["A1"].address[1] is shorter than sheet["A1"].address.split("$")[1] but it is not obvious that one can do that.

If you have the column index, it's usually easier to just work with indices all the way instead of doing cumbersome conversions and string concatenations:

col_ix = ws["A1"].expand().last_cell.offset(row_offset=0, column_offset=1).column
sheet.range((2, 1), (5, col_ix))  # 1-based indexing

For 0-based indexing, use Python's slicing:

sheet[start_row_ix:end_row_ix, start_col_ix:end_col_ix]

Both ways to specify ranges (with indices and with letters) are perfectly fine, there is no way that is superior to the other.
The main thing is that it is accommodating to give people the freedom of choice between column numbers and column letters. Some prefer column numbers, others prefer column letters for easier cognitive handling. Since Microsoft Excel uses letters for numbering columns (for example, demonstrable with the command .address), it seems obvious to offer this letter notation also in .column of xlwings.

Here a less cumbersome version of my last example:

col_letter = ws["A1"].expand().last_cell.offset(row_offset=0, column_offset=1).address.split("$")[1]
row_number = 5

ws.range(f"A2:{col_letter}{row_number}")