Is this a correct way to generate an exception when getting a wrong parameter

Mark Lawrence breamoreboy at yahoo.co.uk
Wed Aug 12 06:43:44 EDT 2015


On 12/08/2015 10:33, Peter Otten wrote:
> Cecil Westerhof wrote:
>
>> I have:
>> ========================================================================
>> accepted_params     = {
>>      'pcpu',
>>      'rss',
>>      'size',
>>      'time',
>>      'vsize',
>> }
>> ========================================================================
>>
>> Later I use:
>> ========================================================================
>> if (to_check != 'all') and not(to_check in accepted_params):
>>      raise Exception('Used illegal parameter: {0}.\n'
>>                      'Accepted ones: {1}'
>>                      .format(to_check, sorted(accepted_params)))
>> ========================================================================
>>
>> When using 'all' I want to do the work for all accepted parameters.
>> ;-)
>
> Doesn't that make it an "accepted parameter"? Why not add it to the set?
>
>> Is this a correct way to do this, or is there a better way?
>
> I suppose you do this early in a function? Then at least choose a more
> specific exception (e. g. ValueError).
>
> If this is about commandline arguments -- argparse can handle such
> restrictions:
>
> $ cat demo.py
> import argparse
> parser = argparse.ArgumentParser()
> parser.add_argument("--check", choices=["pcpu", "rss"], default="all")
> print(parser.parse_args().check)
> $ python3 demo.py
> all
> $ python3 demo.py --check rss
> rss
> $ python3 demo.py --check ssr
> usage: demo.py [-h] [--check {pcpu,rss}]
> demo.py: error: argument --check: invalid choice: 'ssr' (choose from 'pcpu',
> 'rss')
>
>

The wonderful http://docopt.org/ makes this type of thing a piece of 
cake.  I believe there's a newer library that's equivalent in 
functionality to docopt but I can never remember the name of it, anybody?

-- 
My fellow Pythonistas, ask not what our language can do for you, ask
what you can do for our language.

Mark Lawrence




More information about the Python-list mailing list