Comparison problem

[Peter Hansen]

Tim Henderson wrote:
> peter

(Thanks for clarifying to whom you were responding... I saw the other 
post but wouldn't have responded since it didn't seem to be in response 
to one of mine. :-) )

> would not the more correct way to do this be short circuit
> evaluation. somthing along lines of
> 
> if (len(item) > 0) and (item[0] == '-'): pass

This is "correct" only in a narrow sense.  I wouldn't call it "correct 
Python" if I saw it in code around here.  Most likely (since this always 
depends on context), I would replace it with this:

  if item and item[0] == '-':
      pass

I think it's probably unlikely that the first test would be needed only 
on that one line, so it's even more likely that "if item" test would be 
done first, by itself, and then other tests on item[0] would be done.

If this was code that needed high performance (i.e. had been detected 
with proper profiling as a serious bottleneck) then the slicing approach 
would be better (as Fredrik demonstrated).

Hmm... just realized we're going way off track, since we actually have 
the original code here and don't need to comment on hypotheticals.

It appears to me that the original input is being treated as some sort 
of fixed-length record, with certain characters and text in predefined 
positions, specified by index number.  With that in mind, I'd actually 
say the original code is just fine with "if item[0:1] == '-'" and given 
the context it is pretty readable (ignoring the awful formatting with no 
whitespace around operators) and "correct".  Using ".startswith()" might 
have been equally appropriate, but without knowing more detail of the 
input data I couldn't say.

-Peter