`get_savwriter_float_format` is not always calculating the correct length of the int part of a float. This happens because Series.max doesn't return the numeric max when applied on a series with dtype=object. https://github.com/Quantipy/quantipy/blob/82ef628d089ea9dbc47c91e842e50dd9e87bb29e/quantipy/core/tools/dp/spss/writer.py#L133