捕获异常后继续循环

Question

I am reading a large (2Gb) file via csv.DictReader (for what it matters). Somewhere deep in the file there is an issue with the coding and I get a UnicodeDecodeError . Looking at the error message I see that the error is raised in the implicit __next__ in my for loop.

Code stub would look at follows:

import csv

with open("myfile.csv", newline="") as f:
    c = csv.DictReader(f)
    for line in c: # here the error is happening
       pass

I want to use a try-except pattern to catch any reading error, logging a meaningful message and continue reading through the file.

How could I achieve that? I cannot use continue outside the loop (ie in the except block) so I guess I would need to rewrite the for loop in order not to use the implicit form but an explicit one, but as I am fairly new to python, I do not know how to do that in the most pythonic way.

---

To simulate the error look at the following toy example:

class MyIterator:
    def __init__(self, mclass):
        self._mclass = mclass
        self._index = 0

    def __next__(self):
        if self._index == 3:
            # simulate an error in a file read
            self._index += 1
            raise Exception("simulated error"
        elif self._index < 5:
            self._index += 1
            return self._index - 1
        # End of Iteration
        raise StopIteration


class MyClass:
    def __iter__(self):
        return MyIterator(self)

obj = MyClass()
try:
    for result in obj:
        print(result)
except Exception as e:
    print("Exception covered")
    ## does not work for obvious reasons:
    # continue

Answer 1

只是为了记录，而不是试图在next上捕获错误，您可以将errors参数传递给open ，这决定了如何处理编解码器错误。

Answer 2

Wrap your object in an iter and call next on it in a while loop

done = False
obj = iter(MyClass())
while not done:
    try:
        data = next(obj)
    except StopIteration:
        done = True
    except Exception:
         print("process error here and continue")

Answer 3

What python version are u using? Also why not using yield ( see enter link description here )

You can return instead of raise. If you know if type(variable) is Exception you cand process information.