如何在异常中打印有问题的行

Question

我在我的代码中使用Pandas,并使用下面的代码操作一些数据框:

def exit(exitCode, errorMsg=None, monitor_thread=None):
   if exitCode is not None:
       logger.error(errorMsg)
       logger.error(traceback.format_exc())
   sys.exit(exitCode)

try:
    results['name'] = results['name'].astype(float).astype(int)
except Exception as e:  # catch *all* exceptions
    exit("Failed to convert name" + str(e), 8)

我收到此错误:

Traceback(最近一次调用最后一次):

文件"DIST\FDmanInfo.py",第206行,in

在主文件中输入"DIST\FDmanInfo.py",第191行

文件"\ local\Python\lib\site-packages\pandas\core\generic.py",第3054行,在astype raise_on_error = raise_on_error,**kwargs)

文件"\ local\Python\lib\site-packages\pandas\core\internals.py",第3189行,在astype中

return self.apply('astype',dtype = dtype,**kwargs)

文件"\ local\Python\lib\site-packages\pandas\core\internals.py",第3056行,在apply applied = getattr(b,f)(**kwargs)

文件 "\本地\ Python的\ LIB \站点包\大熊猫\核心\ internals.py",线路461,在astype值=值,**kwargs)

文件"\ local\Python\lib\site-packages\pandas\core\internals.py",第504行,_astype values = _astype_nansafe(values.ravel(),dtype,copy = True)

文件"\ local\Python\lib\site-packages\pandas\types\cast.py",第531行,在_astype_nansafe中引发ValueError('无法将NA转换为整数')ValueError:无法将NA转换为整数

我想知道哪一行正在解决这个问题,但是我无法在任何地方找到异常打印它(行)的日志,就像输入数据帧一样:

name

0 1.0

1 NaN

2 3.0

3 4.0

我希望它在它失败的行上打印,在这种情况下"1 NaN",它甚至可能吗？

谢谢,

Answer 1

您可以使用内置的python库,回溯

import traceback
try:
    results['name'] = results['name'].astype(float).astype(int)
except Exception as e:  # catch *all* exceptions
    tb = traceback.format_exc()
    print(tb)
    exit("Failed to convert name" + str(e), 8)

def exit(exitCode, errorMsg=None, monitor_thread=None):
   if exitCode is not None:
       logger.error(errorMsg)
       logger.error(traceback.format_exc())
   sys.exit(exitCode)

更新 检查回溯如何工作的简单示例:

import traceback

try:
    a = [] + None
except Exception as e:
    tb = traceback.format_exc()
    print(tb)
    exit("Failed to convert name" + str(e))


def exit(exitCode):
    print(exitCode)
    print(traceback.format_exc())