Python efficient exception handling

Question

I'm trying to figure out if there's a more pythonic way to do a specific error handling task. I want to catch multiple specific custom exceptions and do something specific depending on which exception is caught (like use a specific custom exit code, or add a specific log message, etc...). However, I also want, if any exception is raised, to do send an email saying the script didn't complete successfully, had exit code __, the log is at C:\foo\bar.txt

I know I could include everything I want to do in each except, such as:

try:
    do_something()
except CustomError1:
    exitcode = 1
    send_error_email(exitcode)
    sys.exit(exitcode)
except CustomError2:
    exitcode = 2
    send_error_email(exitcode)
    sys.exit(exitcode)

But I'm wondering if there's a more pythonic or more efficient way of doing it. I'm imagining something like

try:
    do_something()
except CustomError1:
    exitcode = 1
except CustomError2:
    exitcode = 2
except ParrentCustomErrorClass:
    send_error_email()
    sys.exit(exitcode)

If it matters, I'm currently stuck with python 2.7, but need to be able to port solution to 3.x when 3rd party applications allow it.

modesitt · Accepted Answer · 2020-04-27 22:08:31Z

3

Two idioms I have seen are:

With finally

exit_code = None
try:
   do_something()
except CustomError1:
    exit_code = 1
except CustomError2:
    exit_code = 2
finally:
    if exit_code is not None:
        send_error_mail(exit_code)

With "dictionary dispatch"

try:
    do_something()
except Exception as e:
    code = {
        CustomError1: 1
        CustomError2: 2
    }.get(type(e))
    if code is None:
        raise
    send_error_email(code)

(or if you wanna go full PEP572),

    ...
    if (code := {
        CustomError1: 1
        CustomError2: 2
    }.get(type(e))) is not None:
         send_error_email(code)
    raise

If you control your exceptions however (and they are not library exceptions) - I do like @scnerd's solution too. If you don't - and want to catch ValueErrors or TypeErrors or whatever then one of these is the way to go.

edited Apr 27, 2020 at 22:08

answered Apr 27, 2020 at 22:02

modesitt

7,2492 gold badges38 silver badges66 bronze badges

Sign up to request clarification or add additional context in comments.

2 Comments

mapf Over a year ago

I had a very similar idea about using a dict to store the Exceptions and was about to post it when I saw your answer. I didn't know about these recommended ways. Can you explain why it is better to define that dict again any time an exception is handled? I would have declared a public dictionary errors once containing all the exceptions as keys like you, and whenever an exception had to be handled, did something like except tuple(errors.keys()) as err: ...code = errors.get(type(err)). It would be less code but probably not as fast?

modesitt Over a year ago

It will be just as fast - marginally different if anything. It just depends if this is a one-off kind of thing or you want to be doing this commonly - in which case your way is better :)

scnerd · Accepted Answer · 2020-04-27 22:02:17Z

2

Can you just define the exit codes in the exception types themselves?

class ParentCustomErrorClass(Exception):
    exitcode = 999

    def send_error_email(self):
        raise NotImplementedError()

    def die_horribly(self):
        import sys
        self.send_error_email()
        sys.exit(self.exitcode)

class CustomError1(ParentCustomErrorClass):
    exitcode = 1

class CustomError2(ParentCustomErrorClass):
    exitcode = 2

try:
    do_something()
except ParentCustomErrorClass as ex:
    ex.die_horribly()

answered Apr 27, 2020 at 22:02

scnerd

6,1452 gold badges24 silver badges38 bronze badges

1 Comment

John Over a year ago

I'd never really though of adding error handling into an error class before (sounds stupid, but didn't really think of them as a normal class), so thank you for the insight and will possibly use this in future. For the current project, unfortunately, I don't have control over the error classes however as I'm catching from 3rd party modules.

Collectives™ on Stack Overflow

Python efficient exception handling

2 Answers 2

2 Comments

1 Comment

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

2 Answers 2

2 Comments

1 Comment

Your Answer

Sign up or log in

Post as a guest

Related