writing datasets derived from netCDF4 with compression fails

[matt-long]

When I read a file with netCDF4 compression into a Dataset, a subsequent call to write the dataset using to_netcdf fails.

For instance, using data from the POP model, I can convert output to netCDF4 using NCO

$ ncks --netcdf4 --deflate 1 $file nc4-test.nc

Then in Python:

ds = xr.open_dataset('nc4-test.nc',decode_times=False,decode_coords=False)
ds.to_netcdf('test-out.nc')

The write fails with:

  File "netCDF4/_netCDF4.pyx", line 2263, in netCDF4._netCDF4.Dataset.createVariable (netCDF4/_netCDF4.c:18764)
  File "netCDF4/_netCDF4.pyx", line 3235, in netCDF4._netCDF4.Variable.__init__ (netCDF4/_netCDF4.c:31564)
RuntimeError: NetCDF: Bad chunk sizes.

If I include format = NETCDF3_64BIT, the write completes.

This seems like a bug.

Example dataset: ftp://ftp.ucar.edu/pub/cgd/mclong/misc/nc4-test.nc

[shoyer]

Thanks for the report, I can reproduce this behavior, so I'm marking this as a bug. For completeness I've attached your file below:
test-out.nc.zip

[shoyer]

This looks related to #1225

[stale]

In order to maintain a list of currently relevant issues, we mark issues as stale after a period of inactivity

If this issue remains relevant, please comment here or remove the stale label; otherwise it will be marked as closed automatically

[shoyer]

This works now