issue_comments

I try concatenating on an existing axis, the 'time' axis. I uploaded a couple of files here. It's just easier and you can try experimenting.

Well, I'm not sure if it's a bug, I would say it's more like a missing feature... in my case, each netCDF file has a different mean_height_agl coordinate, that is, they have the same length (it's 1D), but different values in each file. I can understand why it can't concatenate, but I would argue that a better way to handle this is to create a dummy coordinate (as I did) and replace the troublesome coordinate with that dummy coordinate.

I get this with open_mfdataset:

Traceback (most recent call last): File "box.py", line 59, in <module> concat_dim='time')) File "/ichec/home/users/razvan/.local/lib/python3.4/site-packages/xray/backends/api.py", line 205, in open_mfdataset combined = auto_combine(datasets, concat_dim=concat_dim) File "/ichec/home/users/razvan/.local/lib/python3.4/site-packages/xray/core/alignment.py", line 352, in auto_combine concatenated = [_auto_concat(ds, dim=concat_dim) for ds in grouped] File "/ichec/home/users/razvan/.local/lib/python3.4/site-packages/xray/core/alignment.py", line 352, in <listcomp> concatenated = [_auto_concat(ds, dim=concat_dim) for ds in grouped] File "/ichec/home/users/razvan/.local/lib/python3.4/site-packages/xray/core/alignment.py", line 303, in _auto_concat return concat(datasets, dim=dim) File "/ichec/home/users/razvan/.local/lib/python3.4/site-packages/xray/core/alignment.py", line 278, in concat return cls._concat(objs, dim, indexers, mode, concat_over, compat) File "/ichec/home/users/razvan/.local/lib/python3.4/site-packages/xray/core/dataset.py", line 1712, in _concat 'variable %r not %s across datasets' % (k, verb)) ValueError: variable 'mean_height_agl' not equal across datasets

So I don't know if this is what you're asking for (I only have one dataset with this problem) but here's how it looks like:

<xray.Dataset> Dimensions: (agl: 50, lat: 39, lon: 59, time: 192) Coordinates: * lon (lon) float64 -29.0 -28.0 -27.0 -26.0 -25.0 -24.0 ... * lat (lat) float64 32.0 33.0 34.0 35.0 36.0 37.0 38.0 39.0 ... * agl (agl) int64 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 ... * time (time) datetime64[ns] 2011-05-21T13:00:00 ... mean_height_agl (time, agl) float64 28.28 97.21 191.1 310.7 460.9 ... Data variables: so2_concentration (time, agl, lat, lon) float64 3.199e-13 3.199e-13 ... ash_wetdep (time, lat, lon) float64 0.0 0.0 0.0 0.0 0.0 0.0 0.0 ... ash_concentration (time, agl, lat, lon) float64 9.583e-16 9.581e-16 ... ash_mass_loading (time, lat, lon) float64 1.091e-11 1.091e-11 ... so2_mass_loading (time, lat, lon) float64 2.602e-09 2.602e-09 ... ash_drydep (time, lat, lon) float64 4.086e-10 4.084e-10 4.08e-10 ... Attributes:

This is read in with get_ds from above. I wouldn't be able to read it normally with xray.open_mfdataset because it would give me 'Variable mean_height_agl not equal across datasets'. But 'mean_height_agl' is indeed a coordinate per individual file, so I had to create the dummy 'agl' coordinate and convert 'mean_height_agl' to a variable basically. This way I can still treat the data as being part of one file.