Distinguish empty-string cells from blank cells #260

to_python() and iter_rows() return '' for both a blank cell and a cell holding an empty string, so callers can't tell them apart. That matters when importing data where an empty string and a missing value mean different things (e.g. mapping blanks to NULL but keeping '').

import xlwt
from python_calamine import CalamineWorkbook

book = xlwt.Workbook()
sheet = book.add_sheet("s")
sheet.write(0, 0, xlwt.Formula('""'))  # A1: formula whose result is an empty string
sheet.write(0, 1, "x")
sheet.write(1, 1, "x")                 # A2: no cell at all
book.save("t.xls")

print(CalamineWorkbook.from_path("t.xls").get_sheet_by_index(0).to_python())
# [['', 'x'], ['', 'x']]

Tested on 0.9.1. xlrd reports A1 as XL_CELL_TEXT and A2 as XL_CELL_EMPTY.

The calamine crate does keep the difference in its XLS reader (Data::String("") vs an absent cell, in 0.37.0). It seems to be lost in CellValue::Empty => "" in src/types/cell.rs.

Since changing the default would affect existing callers, could this be opt-in, for example a keyword on to_python() / iter_rows() that returns None for blank cells? This is close to #105 / #258 on error cells, which are also indistinguishable from blanks, so one typed-value approach might cover both.