同时处理“,”和“.”的方法作为小数点分隔符

Question

我有一个程序，用户可以复制 Excel 表格并将其粘贴到程序本身上进行数据分析。该表允许的唯一值是浮点数。问题是有 4 种不同的方式来编写浮点数，它们是：

x = 123456.789
y = 123456,789
a = 123.456,789
b = 123,456.789

我需要代码将所有 4 个解释为：

123456.789

最好的方法是什么，才能在 Windows、Linux 和 Mac 上运行？

完整功能如下：

# Copies csv from clipboard to a pandas dataframe
clipboard_dataframe = pd.read_clipboard()
# Get row and column count
paste_rows = len(clipboard_dataframe.index)
paste_columns = len(clipboard_dataframe.columns)
# Create the table to show the values
paste_janela = QTableWidget(paste_rows, paste_columns)
# Define the labels for headers and indexes
paste_janela.setHorizontalHeaderLabels(clipboard_dataframe.columns)
paste_janela.setVerticalHeaderLabels(clipboard_dataframe.index)
# Populate the table with the proper values
for x in range(paste_janela.rowCount()):
    for y in range(paste_janela.columnCount()):
        # Error handling in case the cell value isn't a float
        if not isinstance(clipboard_dataframe.iat[x,y], numbers.Number):
            error_message = QMessageBox.critical(self.janela_principal, 
                                                    "Erro de importação de dados", 
                                                    "Houe um erro na importação de dados de tabela. \nOs únicos valores aceitos são números reais")
            raise ValueError("Valor inválido foi encontrado na tabela. Só é aceito números reais nas tabelas")
        table_value = str(clipboard_dataframe.iat[x,y])
        table_item = QTableWidgetItem(table_value)
        paste_janela.setItem(x, y, table_item)
# Pass the table to the MDI Area, turning it into a subwindow in the process
self.sandbox_mdiarea.addSubWindow(paste_janela)
# Needed to load the window, otherwise it will be hidden as default
paste_janela.show()

Answer 1

这是一个适用于您的测试用例的解决方案，还有我想出的几个解决方案。最大的问题是像

123,456

这样的数字，不清楚它应该是123456还是123.456。

基本的解决方案是检查它是否有小数点或句点，或者两者都有，然后调整字符串以适合内置的

float

函数。假设始终提供小数除法器，我相信这里有一个满足您要求的函数：

def read_float(floatstr: str) -> float:
    period_count = floatstr.count('.')
    comma_count = floatstr.count(',')

    if period_count > 1 and comma_count > 1:
        raise ValueError(f'Unable to read float with multiple of both period and comma: {floatstr}')

    if period_count == 0 and comma_count == 0:
        return float(floatstr)
    if period_count == 0:
        return float(floatstr.replace(',', '.'))
    if comma_count == 0:
        return float(floatstr)

    period_first = floatstr.find('.')
    comma_first = floatstr.find(',')

    if period_first < comma_first:
        return float(floatstr.replace('.', '').replace(',', '.'))
    return float(floatstr.replace(',', ''))


def _main():
    import numpy as np

    test_points = [
        ('123', 123),
        ('1.25', 1.25),
        ('1234.56', 1234.56),
        ('1,234.56', 1234.56),
        ('1,234,567.89', 1234567.89),
        ('1234567.89', 1234567.89),
        ('1,25', 1.25),
        ('1234,56', 1234.56),
        ('1.234,56', 1234.56),
        ('1.234.567,89', 1234567.89),
        ('1234567,89', 1234567.89),
        ('123,456', 123.456)  # this is problematic - no way to tell if 123456 or 123.456
    ]

    for teststr, testval in test_points:
        output = read_float(teststr)
        print(f'{teststr:>12} - {"PASS" if np.equal(testval, output) else "FAIL"}')


if __name__ == '__main__':
    _main()

给出以下输出：

         123 - PASS
        1.25 - PASS
     1234.56 - PASS
    1,234.56 - PASS
1,234,567.89 - PASS
  1234567.89 - PASS
        1,25 - PASS
     1234,56 - PASS
    1.234,56 - PASS
1.234.567,89 - PASS
  1234567,89 - PASS
     123,456 - PASS

我确信存在使用一些本地化库的更简单的解决方案，但我是 找不到它。

如果您有任何疑问，请告诉我。如果您有一个测试点，此功能不起作用，请告诉我，我会考虑更新它。