Open rcpeene opened 3 months ago
@rcpeene see this issue: https://github.com/hdmf-dev/hdmf-zarr/issues/179
The solution mentioned there seems to work. Namely, passing write_args={'link_data': False} to the export call. It is not clear to me if this is intended behavior or merely a temporary solution. The electrodes table and intervals table are not information that is supposed to belong in a separate file. In other words, I think they should be copied over with the default export behavior.
@rcpeene We are working on a fix #194. Ideally rolled out by the end of the week.
What happened?
I am trying to modify an existing zarr nwb file in memory and export a copy with those modifications. Specifically, I am trying to append columns to the nwb.electrodes table. More details discussed here: https://github.com/NeurodataWithoutBorders/helpdesk/discussions/86
After running code similar to the snippet below, the resulting exported NWB does not contain any of the original columns from nwb.electrodes except 'group'. It also includes the columns I have added (x, y, and z). Printing the electrodes table in memory after reading, before exporting, does show all columns and their names.
I examined the nwb.intervals tables as well and they also are missing all columns in the exported nwb.
Steps to Reproduce
Traceback
Operating System
Linux
Python Executable
Conda
Python Version
3.9
Package Versions
hdmf-zarr==0.8.0
Code of Conduct