使用 Python 读取 .txt 或 .csv 格式的 .xlsx

Question

Is there a way to read a .xlsx file in .txt or .csv format with Python? Looking for a way to read an .xlsx file while preserving number formatting (eg, $45.890924). Searched around and could not find a viable module, and creating a style converter would be next to impossible with my Python skill level.

A few helpful notes, Pandas would not be an option because it automatically wipes the number formatting, and I cannot classify the column's format in advance since one column can contain 20+ different number formats.

Answer 1

openpyxl stores the content of the cell in value and the formatting in number_format (and in a few other properties for alignment, color, font, border, etc). So you could interpret the Excel format code and translate it to Python format - but

1. of course a few format properties do not make sense in CSV: eg you cannot make negative numbers red in a CSV
2. While Excel format codes for dates and times are relatively easy to handle, those for numbers can be very tricky to decipher. As an example this is the standard currency format string for Euro:

'_-* #,##0.00\ [$€-410]_-;\-* #,##0.00\ [$€-410]_-;_-* "-"??\ [$€-410]_-;_-@_-'

All that said, making a translator is not impossible. Below is a simple function to translate Excel date format strings to Python's strftime() directives.

def date_xl2py(dt, xlcode):
    xl2py = {
        'yy' : '%y',
        'yyyy' : '%Y',
        'm' : '%m', ##always zero-padded
        'mm' : '%m',
        'mmm' : '%b',
        'mmmm' : '%B',
        'mmmmm' : '%b', ##no single letter form
        'd' : '%d', ##always zero-padded
        'dd' : '%d',
        'ddd' : '%a',
        'dddd' : '%A',
        '%' : '%%' ##escape the % char
        }
    pycode = []
    for xlpart in findall(r'[d|m|y|h|s]+|.|(".+")', xlcode):
        if xlpart in xl2py:
            pycode.append(xl2py[xlpart])
        else:
            pycode.append(xlpart)
    return ''.join(pycode)

dt = datetime(2022,7,12,15,56)
dt.strftime(date_xl2py(dt, 'ddd, mmmm dd, yyyy'))          
'Tue, July 12, 2022'

Please note, I didn't take into account the specification of a locale.

Also, Excel offers three (rather useless) date formatting options that are not available in Python (see comments in the code): I just mapped them to the most similar option available.

And finally, if you were to add time formats, you would need to handle the fact that "mm" may be months or minutes in Excel, and select the right option based on context.