How to clear the exception state in Python

Question

I have a try/except block around a call to an API. It appears to me that once I get an exception, all valid try cases after that exception will see the same exception. The only way I have been able to get it to work is to re-start my Python script. I googled around and found PyErr_clear() but that is for the C-API. Is there something I can call from plain-old Python that will clear the exception state?

Here is the basic idea of what I am doing:

def get_and_print_data(key):
   try:
      data = get_data_from_some_3rd_party_api(key)
   except Exception as ex:
      print("Failed to get data for",key,": ",str(ex))
      return

   print("data=",data)

Then in main I have

get_and_print_data("Valid Key")     ## This works
get_and_print_data("INvalid Key")   ## This gets exception, as expected
get_and_print_data("Valid Key")     ## This and all subsequent calls to get_and_print_data() fail with the same exception.

Can you show your code? I don't know what you mean by "all valid try cases after that exception will see the same exception". — erip
– erip, Commented Oct 5, 2018 at 16:40
The question doesn't make sense without code. It's not a feature of python, it's something you've implemented in code — roganjosh
– roganjosh, Commented Oct 5, 2018 at 16:42
Honestly that seems like a problem with get_data_from_some_3rd_party_api(key) failing after receiving an invalid key. I don't expect that behavior from your try/except. — Jenner Felton
– Jenner Felton, Commented Oct 5, 2018 at 16:48
@Jenner Felton - I was thinking that perhaps something in the 3rd party api needs to be cleared, and that may be so; but then I saw this docs.python.org/3/c-api/exceptions.html which seems to indicate that once set, the exception state stays set until cleared. I have contacted the owner of the API, but thought I would try here as well (while waiting for a response) in case this is a language feature (I'm a C/C++ coder; new to Python). — Daniel Goldfarb
– Daniel Goldfarb, Commented Oct 5, 2018 at 16:50
Can I ask which 3rd party API you are using? I'm pretty convinced it's an issue there. — Jenner Felton
– Jenner Felton, Commented Oct 5, 2018 at 16:54

Jenner Felton · Accepted Answer · 2018-10-05 16:53:38Z

1

As an example of why I think it's the 3rd party API that is having issues:

def get_data_from_some_3rd_party_api(key):
    if key == "Valid Key":
        return "This is some good data."
    else:
        raise ValueError("Invalid Key")

def get_and_print_data(key):
   try:
      data = get_data_from_some_3rd_party_api(key)
   except Exception as ex:
      print("Failed to get data for",key,": ",str(ex))
      return
   print("data=",data)

get_and_print_data("Valid Key")     ## This works
get_and_print_data("INvalid Key")   ## This gets exception, as expected
get_and_print_data("Valid Key")     ## This works

Try running this locally, and you should see that the subsequent valid keys still work.

answered Oct 5, 2018 at 16:53

Jenner Felton

8171 gold badge10 silver badges18 bronze badges

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

5 Comments

Daniel Goldfarb Over a year ago

Thanks. Ran it. That's very convincing. Guess I'll have to wait to hear back from the 3rd party.

Jenner Felton Over a year ago

@DanielGoldfarb Is the 3rd party an open source API with documentation? Happy to help there if it is.

Daniel Goldfarb Over a year ago

I figured out the problem, thanks to your proof that the issue is definitely not Python. The api I am calling is get_daily() from alpha_vantage (github.com/RomelTorres/alpha_vantage) which is a Python wrapper around the alphavantage.co web API. The problem is the free version of the web API is limited to 5 requests per minute; and the Python wrapper defaults to 5 retries on a failure, so when sent an invalid key it burned up my 5 requests per minute. When I set retries=0 everything worked fine. Thanks for your help!

Jenner Felton Over a year ago

@DanielGoldfarb Nice sleuthing and congrats on solving your problem! Glad you figured it out!

Linas Over a year ago

Although you blame the 3rd Party, this also happens if you just use the plain-old C bindings to python. Calling PyErr_Clear() does not seem to actually clear the error. Investigating now.

Collectives™ on Stack Overflow

How to clear the exception state in Python

1 Answer 1

5 Comments

Hot Network Questions

Collectives™ on Stack Overflow

1 Answer 1

5 Comments

Related