Ternary Operator

[Terry Reedy]

"Stefan Kirchberg" <sk at gluit.de> wrote in message
news:3B054397.71A1A29F at gluit.de...
> Hi,
>
> I just read section 4.16 of The Whole Python FAQ
> (http://www.python.org/doc/FAQ.html#4.16). It states that as an
> equivalent to C/C++'s
>         a? b : c
> expression, the Python expression
>         a and b or c
> has a flaw if b evaluates to 0 (false)

However, if one knows that b cannot be false, this works wonderfully.

> and the expression
>         (a and [b] or [c])[0]
> being a possible solution. Unfortunately, the latter one seems
> kind of obfuscated to me...
>
> Now, how about
>         [b,c][not a]

Your alternative requires that both b and c be computed, regardless of the
value of a, which is a flaw when either b or c will raise an exception if a
has the wrong value.  Guarding the computation of b by precondition a is a
common use of ternarys.  Consider
  y and x/y or None # or  (y and (x/y,) or (None,))[0] # versus
(x/y,None)[not y]

TJ Reedy