Re: Search pattern for non-ASCII alphabetic characters

Janis Papanagnou wrote:
> Hermann Peifer wrote:
>> Hi,
>>
>> Occasionally, I'd like to search for non-ASCII alphabetic characters 
>> in UTF-8 encoded text documents.
>>
>> In the absence of an appropriate character class (at least I wouldn't 
>> know of any), I do something like:
>>
>> awk '/[ÀÁÂÃÄÅ ...and so on... ŸŹźŻżŽž]/{ action }'
>>
>> This is perhaps not the smartest solution. Any better idea?
>>
>> TIA. Hermann
> 
> I can't tell if it is a smarter solution but you could use the inverse
> logic based on the existing character classes...
> 
>   LANG=C  awk '/[^[:alnum:][:punct:][:blank:][:cntrl:]]/'
> 
> (Note: there's also the ANSI character class [:ascii:] but my GNU awk
> seems to not support it.)
> 
> Janis


Thanks for the hint. This pattern also finds:  N°1
Which does not exactly contain a non-ASCII *alphabetic* character, but 
it's still better than the long character list I was using. I can filter 
out some false positives.

Hermann