python - Python 中 try-catch-raise 异常子句的两次运行行为

Question

在下面的最小示例中，我用我自己的方法 safe_apply 包装了 pandas.DataFrame.apply 方法。此方法在两个方面不同于普通的 apply:

1. 它不是在 NDFrame 中返回值，而是将它们附加到列表 results 中。
2. 如果异常被捕获，results 会在错误出现和执行终止之前被打印出来。

代码如下:

from pandas.core.frame import DataFrame, Series
import pandas as pd

def safe_apply(df, func, **kwargs):
    results = []

    def new_func(srs):
        try:
            results.append(func(srs))
        except Exception as e:
            print(results)
            raise

    df.apply(new_func, **kwargs)
    return results

DataFrame.safe_apply = safe_apply

def f(srs):
    if (pd.notnull(srs['lat'])) & (pd.notnull(srs['long'])):
        return srs['lat'] + srs['long']
    else:
        raise ValueError


ex = pd.DataFrame({'lat': [1, 2, None], 'long': [1, 2, None]}, index=['A', 'B', 'C'])
ex.safe_apply(f, axis='columns')

当我执行这个时，我除了将函数 f 应用于 ex 的前两行然后在第三行失败的结果:a print 来自 [2.0, 4.0]，后跟 ValueError

相反，我得到了一个ValueError 和两行 输出:

[2.0, 4.0]
[2.0, 4.0, 2.0, 4.0]
---------------------------------------------------------------------------
ValueError                                Traceback (most recent call last)
<ipython-input-1-382c07ef0919> in <module>()
     26         raise ValueError
     27 
---> 28 ex.safe_apply(f, axis='columns')

<ipython-input-1-382c07ef0919> in safe_apply(df, func, **kwargs)
     12             raise
     13 
---> 14     df.apply(new_func, **kwargs)
     15 
     16     return results

C:\Users\Alex\Anaconda3\lib\site-packages\pandas\core\frame.py in apply(self, func, axis, broadcast, raw, reduce, args, **kwds)
   4059                     if reduce is None:
   4060                         reduce = True
-> 4061                     return self._apply_standard(f, axis, reduce=reduce)
   4062             else:
   4063                 return self._apply_broadcast(f, axis)

C:\Users\Alex\Anaconda3\lib\site-packages\pandas\core\frame.py in _apply_standard(self, func, axis, ignore_failures, reduce)
   4155             try:
   4156                 for i, v in enumerate(series_gen):
-> 4157                     results[i] = func(v)
   4158                     keys.append(v.name)
   4159             except Exception as e:

<ipython-input-1-382c07ef0919> in new_func(srs)
      7     def new_func(srs):
      8         try:
----> 9             results.append(func(srs))
     10         except Exception as e:
     11             print(results)

<ipython-input-1-382c07ef0919> in f(srs)
     24         return srs['lat'] + srs['long']
     25     else:
---> 26         raise ValueError
     27 
     28 ex.safe_apply(f, axis='columns')

ValueError: occurred at index C

意思是代码到达异常子句，打印出结果，然后再次运行该函数而不是抛出错误并停止，以某种方式再次运行该函数(？)。

为什么会这样？

Answer 1

请参阅 docstring 上的注释部分, 具体来说

In the current implementation apply calls func twice on the first column/row to decide whether it can take a fast or slow code path. This can lead to unexpected behavior if func has side-effects, as they will take effect twice for the first column/row.