Creating a conda environment as you suggest, I am fully able to read etc the file, so this solves my issue. Many thanks! Then I guess this means there is some weird issue leading to segfaults on this file with some of the older libnetcdf versions. Closing as using a conda env and a more recent stack fixes things.

I do not provide a dedicated engine in the open_dataset command, so I guess this is using the netcdf4 engine by default?

Depends on the file format, but in general, I'd recommend passing the engine kwarg explicitly.

Ok, thanks, this does crash too on my machine. Then likely something to do with my software stack somewhere, I will try with a new mamba / conda environment and check if this fixes things.

This is what I'm doing: python In [1]: import xarray as xr ...: ...: ds = xr.open_dataset("2018_03_b.nc", decode_times=False) ...: ds["accD"].data Out[1]: array([[-9.625633e+00, -9.587038e+00, -9.601438e+00, ..., -9.543669e+00, -9.519094e+00, -9.509479e+00], [-9.705109e+00, -9.658272e+00, -9.672318e+00, ..., 9.969210e+36, 9.969210e+36, 9.969210e+36]], dtype=float32)

@keewis regarding the engine: I have netcdf4 installed and I do not provide a dedicated engine in the open_dataset command, so I guess this is using the netcdf4 engine by default? Anyway I can run a command to double check and confirm to you? :)

@keewis interesting. Just to be sure: I am able to open the dataset just fine too, the issue arises when trying to actually read the field, i.e.:

python xr_file = xr.open_dataset(input_file, decode_times=False)

is just fine, but

python xr_file["accD"][0, 0:3235893].data

is what segfaults; just to be sure there is no misunderstanding, are you actually able to run the last command without isse? :)

I'm able to open the dataset just fine, so I'd think this has something to do with your environment. Can you try creating a more recent one, using something like: mamba create -n test -c conda-forge python=3.10 xarray netcdf4 ... Also, what engine are you using to read the file? netcdf4?

I got help to extract more information in gdb; converting the ipynb to a py file and running it in gdb context:

```

jupyter nbconvert --to script issue_opening_2018_03_b.ipynb [NbConvertApp] Converting notebook issue_opening_2018_03_b.ipynb to script [NbConvertApp] Writing 1313 bytes to issue_opening_2018_03_b.py gdb --args python3 issue_opening_2018_03_b.py [...] (gdb) run Starting program: /usr/bin/python3 issue_opening_2018_03_b.py [Thread debugging using libthread_db enabled] Using host libthread_db library "/lib/x86_64-linux-gnu/libthread_db.so.1". [...] Thread 1 "python3" received signal SIGSEGV, Segmentation fault. __memmove_sse2_unaligned_erms () at ../sysdeps/x86_64/multiarch/memmove-vec-unaligned-erms.S:314 314 ../sysdeps/x86_64/multiarch/memmove-vec-unaligned-erms.S: No such file or directory. (gdb) bt

0 __memmove_sse2_unaligned_erms () at ../sysdeps/x86_64/multiarch/memmove-vec-unaligned-erms.S:314

1 0x00007ffff6af4bdc in NC4_get_vars () from /home/jrmet/.local/lib/python3.8/site-packages/netCDF4/.libs/libnetcdf-5e98d7e6.so.15.0.0

2 0x00007ffff6af337d in NC4_get_vara () from /home/jrmet/.local/lib/python3.8/site-packages/netCDF4/.libs/libnetcdf-5e98d7e6.so.15.0.0

3 0x00007ffff6a959aa in NC_get_vara () from /home/jrmet/.local/lib/python3.8/site-packages/netCDF4/.libs/libnetcdf-5e98d7e6.so.15.0.0

4 0x00007ffff6a96b9b in nc_get_vara () from /home/jrmet/.local/lib/python3.8/site-packages/netCDF4/.libs/libnetcdf-5e98d7e6.so.15.0.0

5 0x00007ffff6ec24bc in ?? () from /home/jrmet/.local/lib/python3.8/site-packages/netCDF4/_netCDF4.cpython-38-x86_64-linux-gnu.so

6 0x00000000005f5b39 in PyCFunction_Call ()

```

which seems to be originating in libnetcdf?