search

zarr-developers / VirtualiZarr

Create virtual Zarr stores from archival data files using xarray syntax
https://virtualizarr.readthedocs.io/en/stable/api.html
Apache License 2.0
123 stars 24 forks source link

What should nbytes return? #167

Open TomNicholas opened 4 months ago

TomNicholas commented 4 months ago

Xarray uses the optional property .nbytes to indicate the size of wrapped arrays. Currently we don't implement .nbytes on ManifestArrays, so xarray defaults to estimating the size as basically arr.size * arr.dtype.itemsize. I.e. currently it returns what the full size of the dataset would be if you loaded every referenced chunk into memory at once. But does this make sense for an array that can never be loaded into memory?

There is another completely different size to consider - that of the in-memory representation of the references - see discussion in #104. This is a known fixed number, it's not lazy-loading, but it's much smaller than the current nbytes.

This latter number is what's relevant if you're trying to estimate RAM usage whilst manipulating references, so it's possibly related to the __sizeof__ discussion in https://github.com/pydata/xarray/issues/5764.

dcherian commented 2 months ago

But what if you did want to estimate total size of the dataset?

TomNicholas commented 2 months ago

Yes, fair question, that doubt is why I haven't merged #227 yet.

Perhaps we should have two ways to display the size, one normal ds.nbytes and one using the accessor e.g. ds.virtualize.nbytes. I'm not sure whether it would be more intuitive to have the normal ds.nbytes be the actual in-memory size used or the memory that would be taken up by the whole dataset though.

maxrjones commented 1 day ago

+1 on the accessor idea to show the virtual dataset size and maintaining the behavior of the current .nbytes attribute