[Numpy-discussion] byteswap() leaves dtype unchanged
Erin Sheldon
erin.sheldon at gmail.com
Wed May 30 10:21:03 EDT 2007
Matthew, this is a very clear description of all the issues, and I now
see why it can be useful to keep the two methods separate. I think an
update to the doc string for byteswap() with this description would be
useful. Or perhaps a keyword to byteswap() in which one could specify
the behavior regarding the data type.
Erin
On 5/30/07, Matthew Brett <matthew.brett at gmail.com> wrote:
> Hi,
>
> > > This doesn't make any sense. The numbers have changed but the dtype
> > > is now incorrect. If you byteswap and correct the dtype the numbers
> > > have still changed, but you now can actually use the object.
> >
> > By "numbers have still changed" I mean the underlying byte order is
> > still different, but you can now use the object for mathematical
> > operations. My point is that the metadata for the object should be
> > correct after using it's built-in methods.
>
> I think the point is that you can have several different situations
> with byte ordering:
>
> 1) Your data and dtype endianess match, but you want the data swapped
> and the dtype to reflect this
>
> 2) Your data and dtype endianess don't match, and you want to swap the
> data so that they match the dtype
>
> 3) Your data and dtype endianess don't match, and you want to change
> the dtype so that it matches the data.
>
> I guess situation 1 is the one you have, and the way to deal with this
> is, as you say:
>
> other_order_arr = one_order_arr.byteswap()
> other_order_arr.dtype.newbyteorder()
>
> The byteswap method is obviously aimed at situation 2 (the data and
> dtype don't match in endianness, and you want to swap the data).
>
> I can see your point I think, that situation 1 seems to be the more
> common and obvious, and coming at it from outside, you would have
> thought that a.byteswap would change both.
>
> Best,
>
> Matthew
> _______________________________________________
> Numpy-discussion mailing list
> Numpy-discussion at scipy.org
> http://projects.scipy.org/mailman/listinfo/numpy-discussion
>
More information about the NumPy-Discussion
mailing list