[Bug]: Appending to ElectrodesTable in Zarr

[oruebel]

What happened?

This issue was originally raised on PyNWB here NeurodataWithoutBorders/pynwb#2055 (comment) by @FrancescoNegri

I would like to reopen this issue as we switched to the zarr backend @stephprince.

In that case, according to my understanding, there is no maxshape attribute to edit. The append_data function in data_utils.py called by add_electrode works by appending a new row to the data object of type ZarrArray (the dataset to modify). Unluckily, data.read_only is set to True, even if I opened the .nwb file with mode="r+".

I double checked the same ZarrArray by opening it via the zarr module and read_only == False, making the dataset editable. Is this behaviour expected? Is there a workaround I could exploit to add new electrodes to an existing electrode table?

Thanks in advance to everyone!

Steps to Reproduce

See https://github.com/NeurodataWithoutBorders/pynwb/issues/2055

Traceback

None

Operating System

macOS

Python Version

3.13

Package Versions

No response

[oruebel]

See NeurodataWithoutBorders/pynwb#2055

[oruebel]

Is this behaviour expected? Is there a workaround I could exploit to add new electrodes to an existing electrode table?

@FrancescoNegri As a workaround, you can try to handZarrIO or NWBZarrIO (whichever you are using) an already opened Zarr DirectoryStore object instead of just providing the path. In this way you can control directly how the file is being opened. If that doesn't work, then there may be an issue with how a particular column is being accessed (e.g., if an object is being accessed via a Link it is likely being opened as read only), however, I don't think that should be the case.

[FrancescoNegri]

@oruebel Thanks for your answer.
I tried the following, but the read_only attribute for ElectrodesTable columns (I tested multiple colnames) remains set to True:

import zarr
from hdmf_zarr.nwb import NWBZarrIO

store = zarr.DirectoryStore("path/to/my/file.nwb")
zarr.open(store, mode="r+")

with NWBZarrIO(store, mode="r+") as io:
    nwbfile = io.read()
    print(nwbfile.acquisition["ElectricalSeries"].data.read_only)   # False
    print(nwbfile.electrodes["id"].data.read_only)                  # True