RFC: more generic DataPoint values

[ignatz]

Hi,

I've been toying with the idea of using tstorage for a series of location data, i.e. 3D positions (latitude, longitude, altitude) for different devices.
If I understand correctly, using tstorage I'd have to model the various serious and values belonging to the same "measurement" as something like:

[
(DataPoint(now, 42), Label("id", "device0"), Label("latitude", "")),
(DataPoint(now, 23), Label("id", "device0"), Label("longitude", "")),
...
]

Is my understanding correct? If so and it's probably a big ask, but would it be possible/make-sense to generalize the data representation to e.g. a byte array? In that case, I could store all related measures as a single point. Does the storage format currently require fixed value/data-point sizes?

Thanks!

[swithek]

I also think a more generic/byte array value type would be useful. @nakabonne have you considered that? This would require a new major version to be released but it would be a very nice addition.