shouldn't 'string'.find('ugh') return 0, not -1 ?

[J. Clifford Dyer]

On Wed, Oct 31, 2007 at 03:55:49PM +0100, jelle feringa wrote regarding Re: shouldn't 'string'.find('ugh') return 0, not -1 ?:
> 
>    There is a subtle point though.
> 
>    If the substring is not found '_'.find(' '), will return -1
> 
>    Semanticly, I was expecting the that if the substring was not found,
>    the conditional statement would not be found.
> 
>    However, python evaluates -1 to True, so that is what I do find
>    confusing.
> 
>    So, I was arguing that '_'.find(' ') might return 0, however that is
>    obviously ambigious, since 0 might be an index as well.
> 
> 
> 
>    So, perhaps I should rephrase and ask, why if -1 evaluates to True?
> 
>    I think that's pretty ugly...
> 
> 

It is ugly, but there's no alternative.  In essence, you can't do a truth-test on it, because you're not testing the truth of the find statement.  The statement that you want to test the truth of is s.find(q) >= 0.  In other words, you want to see that the substring was found at a valid (non-negative) location. As someone else pointed out, it would make more sense to use None instead of -1.  You still couldn't use `if s.find(q):` because you'd get a false result for a substring found at the beginning of the string, but the return value would make more intuitive sense.  

Witness also the problem of actually using the result of find:

s[s.find(q)]

will yield a valid part of the string (the last character) when q is not found.  If it evaluated to none, you'd get a respectable TypeError instead.

But you still wouldn't--and never will--be able to say `if s.find(q)`, because valid array indices can be either true or false.

Cheers,
Cliff