The secrets of operator.add

[Paul Prescod]

Jacek Generowicz wrote:
> 

>    ncalls  tottime  percall  cumtime  percall filename:lineno(function)
>     49999    0.840    0.000    0.840    0.000 <string>:1(<lambda>)
>         1    0.720    0.720    1.560    1.560 <string>:1(?)
>         0    0.000             0.000          profile:0(profiler)
>         1    0.040    0.040    1.600    1.600 profile:0(reduce(lambda x,y:x+y, range(N)))
> 
> ...
>    ncalls  tottime  percall  cumtime  percall filename:lineno(function)
>         1    0.060    0.060    0.060    0.060 <string>:1(?)
>         0    0.000             0.000          profile:0(profiler)
>         1    0.000    0.000    0.060    0.060 profile:0(reduce(add, range(N)))
> 
> It looks as if multiple calls to operator.add are avoided somehow
> . . . or am I failing to understand the profiler's output ?

Operator.add is a builtin, not a Python function so calling it is very
quick and it doesn't really show up in the profiler.

> Is there a way of getting similar advantages in
> 
> run ('map(lambda x:x+1, range(N))')  ?

How about:

run('[x+1 for x in range(N)]')

> Is there a Python way of currying add?

lambda is the way to do currying. It may not be as efficient as you like
but that's how it works...

 Paul Prescod