Does it have native bytes support? To me, it's either having fast conversion to `string` or data format with native bytes support.
Sometimes readability is important, sometimes speed takes priority. Even with a good, unified data structure for arrays, indexed arrays, etc., it is always good to know that if it becomes a bottleneck I can substitute array value with its bytes representation with little to none extra work. DG > On 25 Feb 2024, at 21:02, Robert Kern <[email protected]> wrote: > > On Sun, Feb 25, 2024 at 1:52 PM Dom Grigonis <[email protected] > <mailto:[email protected]>> wrote: > Thank you for your answer, > > Yeah, I was unsure if it ever existed in the first place. > > Space is less of an issue and something like `a.data.hex()` would be fine as > long as its speed was on par with `a.tobytes()`. However, it is 10x slower on > my machine. > > This e-mail is pretty much a final check (after a fair bit of research and > attempts) that it can not be done so I can eliminate this possibility as > feasible and concentrate on other options. > > I think that mostly runs the gamut for pure JSON. Consider looking at BJData, > but it's "JSON-based" and not quite pure JSON. > > https://neurojson.org/ <https://neurojson.org/> > > -- > Robert Kern > _______________________________________________ > NumPy-Discussion mailing list -- [email protected] > To unsubscribe send an email to [email protected] > https://mail.python.org/mailman3/lists/numpy-discussion.python.org/ > Member address: [email protected]
_______________________________________________ NumPy-Discussion mailing list -- [email protected] To unsubscribe send an email to [email protected] https://mail.python.org/mailman3/lists/numpy-discussion.python.org/ Member address: [email protected]
