[Numpy-discussion] BOF notes: Fernando's proposal: NumPy ndarray with named axes

[Joshua Holbrook]

On Thu, Jul 8, 2010 at 12:55 PM, Lluís <xscript at gmx.net> wrote:
> Fernando Perez writes:
>> The consensus at the  BoF (not that it means it's set in stone, simply
>> that there was  good chance for back-and-forth on the topic with many
>> voices) was that:
>
>> 1. There are valid use cases for 'integer ticks',  i.e. integers that
>> index arbitrarily into an  array instead of in 0..N-1 fashion.
>
>> 2. That having plain arr[0] give anything but the first element in arr
>> would be way too confusing in practice, and likely to cause too many
>> problems.
>
>> 3. That the  best solution to allow integer ticks while retaining
>> 'normal' indexing semantics for integers would be to have
>
>> arr[int] -> normal indexing
>> arr.somethin[int] -> tick-based indexing, where an int can mean anything.
>
> Thanks for the summary. Then, does it make sense this scheme?
>
>  arr[int] -> normal indexing
>  arr[not int] -> IndexError
>  arr.something[int] -> normal indexing (use str(int) for tick-based indexing)
>  arr.something[not int] -> tick-based indexing
>
> It's not that I'm trying to push for this scheme. In fact, my question is
> whether this would be of any practical use. (e.g., arr.something['1900':-3], or
> arr.named[0,'1900':])
>
> If that is acceptable and performance wouldn't have a huge drop, then I'd rather
> go for:
>
>  arr[int] -> normal indexing
>  arr[not int] -> tick-based indexing
>                  I'm not sure if this was dropped because of the str(int) or
>                  because of performance reasons
>  arr.something[int] -> normal indexing
>  arr.something[not int] -> tick-based indexing
>  arr.named -> no longer necessary
>
> Which has even less "keywords" to remember.
>
>
> Lluis
>
> --
>  "And it's much the same thing with knowledge, for whenever you learn
>  something new, the whole world becomes that much richer."
>  -- The Princess of Pure Reason, as told by Norton Juster in The Phantom
>  Tollbooth
> _______________________________________________
> NumPy-Discussion mailing list
> NumPy-Discussion at scipy.org
> http://mail.scipy.org/mailman/listinfo/numpy-discussion
>

Lluis said:

>  arr[not int] -> tick-based indexing
>                 I'm not sure if this was dropped because of the str(int) or
>                 because of performance reasons

At the BoF, we chose to drop this because we wanted to allow integer
ticks (or implicit type conversion, either way) without the ambiguity
of, "did we mean that in the ndarray sense or in a "tick with the name
'1'" sense? The choice, at the time, had nothing to do with
performance reasons.

In fact, worrying about speed *now* would be a bad idea. We should try
to optimize everything *after* we have an interface that we like.

--Josh