Joe Hamman
Joe Hamman
One path forward could be to raise an error if `open_array` finds an array that does not match the provided parameters (shape, type, etc.). Would that suffice here?
This sounds like an issue with the `RemoteStore` in zarr3. I wonder if something bad is happening because `S` is a single character var name.
Thanks for digging into this @TomAugspurger - @rabernat and I hit something similar a week ago. See https://github.com/zarr-developers/zarr-python/pull/2348 for a zarr-side attempt at working around the fsspec cache issues.
Meta comment - we are working on a fresh Array implementation that covers both V2 and V3 arrays (see #1583 for details). We are actively seeing input/participation from those invested...
@martindurant and others - now that v3 has come together, I'd be very interested to see this move to that branch. Who is interested in trying out an implementation on...
Now would be a great time to start thinking about how this will work with Zarr-Python 3.
@d-v-b - 75% sure this can be closed now that #2442 is in. What do you think?
Nearly all of these tests/stores were removed with 3.0 so I think we can safely close this.
@d-v-b - I think @normanrz's PR (#1857) superseded this effort. Feel free to re-open if I have that wrong.
Proposal from today's dev meeting. We remove `copy` and `copy_all` and push `copy_store` to after 3.0.