What is your issue?

I'm not sure yet if this is a feature or a bug - I would tend to the latter. Apologies if this has been discussed before.

Objects originating from operations such as y = x > 2 are now sharing the same attrs, which leads to things like:

```python import numpy as np import xarray as xr xr.version '2022.6.0'

x = xr.DataArray( 0.1 * np.arange(10), dims=["lat"], coords={"lat": np.arange(10)}, name="sst", ) x.lat.attrs['long_name'] = 'latitude' x.lat.attrs {'long_name': 'latitude'}

y = x > 2 y.lat.attrs {'long_name': 'latitude'}

y.lat.attrs = {} x.lat.attrs # x is changed as well! {} ```

I think this is rather a non-intuitive behavior but I'm happy to discuss!

Consider:

python da = xr.DataArray([1, 2, 3]) print(f'{da[0]}') print(f'{da[0]:d}')

Which outputs:

``` <xarray.DataArray ()> array(1)

TypeError Traceback (most recent call last) <ipython-input-36-9cd7dc76455b> in <module> 1 da = xr.DataArray([1, 2, 3]) 2 print(f'{da[0]}') ----> 3 print(f'{da[0]:d}')

TypeError: unsupported format string passed to DataArray.format ``` And the numpy equivalent:

python da = xr.DataArray([1, 2, 3]).data print(f'{da[0]}') print(f'{da[0]:d}') 1 1

I always found the xarray scalar output to be a bit unfriendly for beginners. In my classes very often scalars are the last output of a computation, and the fact that we can't format the relatively verbose xarray output without resulting to the .data trick is a bit confusing for students (but I agree this is a detail).

Is there a way to get print(f'{da[0]:d}') to work? Thoughts?

This adds some warnings at the right places when rechunking a dataset opened with open_mfdataset (see https://github.com/pangeo-data/rechunker/issues/100#issuecomment-1116189019 for context)

Thanks to @dcherian for the wisdom of the day!

• [x] Closes https://github.com/pydata/xarray/issues/5976
• [x] Tests added
• [x] Passes pre-commit run --all-files
• [ ] User visible changes (including notable bug fixes) are documented in whats-new.rst

This is a first try at formatting dataarray scalars. Here is the current behavior:

```python In [1]: import xarray as xr ...: import numpy as np

In [2]: a = np.array(1) ...: da = xr.DataArray(a)

In [3]: print(a) 1

In [4]: print(da) <xarray.DataArray ()> array(1)

In [5]: print('{}'.format(a)) 1

In [6]: print('{}'.format(da)) <xarray.DataArray ()> array(1)

In [7]: print('{:.3f}'.format(a)) 1.000

In [8]: print('{:.3f}'.format(da)) 1.000

In [9]: a = np.array([1, 2]) ...: da = xr.DataArray(a)

In [10]: print('{}'.format(a)) [1 2]

In [11]: print('{}'.format(da)) <xarray.DataArray (dim_0: 2)> array([1, 2]) Dimensions without coordinates: dim_0

In [12]: print('{:.3f}'.format(a))

TypeError Traceback (most recent call last) <ipython-input-12-c5afc7863e89> in <module> ----> 1 print('{:.3f}'.format(a))

TypeError: unsupported format string passed to numpy.ndarray.format

In [13]: print('{:.3f}'.format(da))

TypeError Traceback (most recent call last) <ipython-input-13-bddebd8462bd> in <module> ----> 1 print('{:.3f}'.format(da))

~/disk/Dropbox/HomeDocs/git/xarray/xarray/core/common.py in format(self, format_spec) 162 return formatting.array_repr(self) 163 # Else why fall back to numpy --> 164 return self.values.format(format_spec) 165 166 def _iter(self: Any) -> Iterator[Any]:

TypeError: unsupported format string passed to numpy.ndarray.format ``` I don't think there is any backwards compatibility issue but lets see if the tests pass

Closes #3080

Some attribute names aren't valid netCDF4 keys. For example:

python import xarray as xr ds = xr.Dataset({'x': ('y', [1, 2, 3], {'CLASS': 'foo'})}) ds.to_netcdf('test.nc')

will fail with:

NetCDF: String match to name in use

This is hard to debug if you don't know which attribute is faulty. With this small change the error should be more informative:

AttributeError: The following exception occurred when attempting to set attribute (CLASS, foo): "NetCDF: String match to name in use"

A very modest contribution... I miss contributing to xarray!

Believe it or not, I did not know about to_pandas until today...

Salem's tests on master fail because I use the lock kwarg to open_dataset, which seems to have disappeared in the backend refactoring.

Should the new open_dataset simply ignore lock, and raise a FutureWarning when used?

This is: - xarray: 0.16.2 - dask: 2.30.0

I'm not sure a github issue is the right place to report this, but I'm not sure where else, so here it is.

I just had two very long weeks of debugging stalled (i.e. "dead") OGGM jobs in a cluster environment. I finally nailed it down to ds.to_netcdf(path) in this situation:

python with xr.open_mfdataset(tmp_paths, combine='nested', concat_dim='rgi_id') as ds: ds.to_netcdf(path) tmp_paths are a few netcdf files (from 2 to about 60). The combined dataset is nothing close to big (a few hundred MB at most).

Most of the time, this command works just fine. But in 30% of the cases, this would just... stop and stall. One or more of the workers would simply stop working without coming back or erroring.

What I can give as additional information: - changing ds.to_netcdf(path) to ds.load().to_netcdf(path) solves the problem - the problem became worse (i.e. more often) when the files to concatenate increased in the number of variables (the final size of the concatenated file doesn't seem to matter at all, it occurs also with files < 1 MB) - I can't reproduce the problem locally. The files are here if someone's interested, but I don't think the files are the issue here. - the files use gzip compression - On cluster, we are dealing with 64 core nodes, which do a lot of work before arriving to these two lines. We use python multiprocessing ourselves before that, create our own pool and use it, etc. But at the moment the job hits these two lines, no other job is running.

Is this is some kind of weird interaction between our own multiprocessing and dask? Is it more an IO problem that occurs only on cluster? I don't know.

I know this is a crappy bug report, but the fact that I lost a lot of time on this recently has gone on my nerves :wink: (I'm mostly angry at myself for taking so long to find out that these two lines were the problem).

In order to make a question out of this crappy report: how can I possibly debug this? I solved my problem now (with ds.load()), but this is not really satisfying. Any tip is appreciated!

cc @TimoRoth our cluster IT whom I annoyed a lot before finding out that the problem was in xarray/dask

Everything is in the title! I think it should, if possible.

The current doc shows bits of code which are now irrelevant thanks to open_mfdataset: http://xarray.pydata.org/en/stable/io.html#id7

On a related note, it would be great to document the bottlenecks in concat / dask and how to overcome them. Related to https://github.com/pydata/xarray/issues/1391, https://github.com/pydata/xarray/issues/1495, and https://github.com/pydata/xarray/issues/1379

Is this a bug? Look the following example:

```python ds = xr.tutorial.load_dataset('air_temperature')

ds.sel(time='2013-01-01T00:00') # works fine [output removed]

ds.sel(time=['2013-01-01T00:00']) # errors Traceback (most recent call last): File "/home/mowglie/.pycharm-community-2017.1/helpers/pydev/_pydevd_bundle/pydevd_exec2.py", line 3, in Exec exec(exp, global_vars, local_vars) File "", line 1, in <module> File "/home/mowglie/Documents/git/xarray-official/xarray/core/dataset.py", line 1206, in sel self, indexers, method=method, tolerance=tolerance File "/home/mowglie/Documents/git/xarray-official/xarray/core/indexing.py", line 290, in remap_label_indexers dim, method, tolerance) File "/home/mowglie/Documents/git/xarray-official/xarray/core/indexing.py", line 229, in convert_label_indexer % index_name) KeyError: "not all values found in index 'time'" ```

Problem

Here is the dump of a NetCDF file (download):

``` netcdf cesm.TREFHT.160001-200512.selection { dimensions: time = UNLIMITED ; // (4872 currently) lat = 3 ; lon = 3 ; nbnd = 2 ; variables: float TREFHT(time, lat, lon) ; TREFHT:units = "K" ; TREFHT:long_name = "Reference height temperature" ; TREFHT:cell_methods = "time: mean" ; double lat(lat) ; lat:long_name = "latitude" ; lat:units = "degrees_north" ; double lon(lon) ; lon:long_name = "longitude" ; lon:units = "degrees_east" ; double time(time) ; time:long_name = "time" ; time:units = "days since 0850-01-01 00:00:00" ; time:calendar = "noleap" ; time:bounds = "time_bnds" ; double time_bnds(time, nbnd) ; time_bnds:long_name = "time interval endpoints" ;

// global attributes: :Conventions = "CF-1.0" ; :source = "CAM" ; ... } ```

When xarray decodes the time coordinates it also deletes the time:units attribute (this kind of makes sense, because the unit has no meaning when the time is converted to a CFTime object):

python import xarray as xr ds = xr.open_dataset(f) ds.time <xarray.DataArray 'time' (time: 4872)> array([cftime.DatetimeNoLeap(1600, 2, 1, 0, 0, 0, 0, 0, 32), cftime.DatetimeNoLeap(1600, 3, 1, 0, 0, 0, 0, 0, 60), cftime.DatetimeNoLeap(1600, 4, 1, 0, 0, 0, 0, 3, 91), ..., cftime.DatetimeNoLeap(2005, 11, 1, 0, 0, 0, 0, 6, 305), cftime.DatetimeNoLeap(2005, 12, 1, 0, 0, 0, 0, 1, 335), cftime.DatetimeNoLeap(2006, 1, 1, 0, 0, 0, 0, 4, 1)], dtype=object) Coordinates: * time (time) object 1600-02-01 00:00:00 ... 2006-01-01 00:00:00 Attributes: long_name: time bounds: time_bnds

The problem is that I have no way to actually decode the time_bnds variable from xarray alone now, because the time_bnds variable doesn't store the time units. First, I thought that my file was not CF compliant but I've looked into the CF conventions and it looks like they are not prescribing that time_bnds should also have a units attribute.

Solution

I actually don't know what we should do here. I see a couple of ways: 1. we don't care and leave it to the user (here: me) to open the file with netCDF4 to decode the time bounds 2. we don't delete the time:units attribute after decoding 3. we start to also decode the time_bnds when available, like we do with time

Thoughts? cc @spencerkclark @jhamman

conda-forge now ships v1.0.1 per default.

After two years of betas and release candidates this is very welcome!

We have very little code specifically handling pre-v1 and post-v1, but we should keep it around for a couple more months.

Will update the CI to reflect this change.

• [x] Closes https://github.com/pydata/xarray/issues/2565
• [x] Tests added
• [x] Fully documented, including whats-new.rst for all changes and api.rst for new API

Not sure if this is the best way to handle it, but it seems to work

(EDIT: sorry for unexpected early posting)

I just wanted to open this issue here, just to see if it has some resonance in other projects.

We are getting new unexpected HDF Errors in our test suite which are definitely due to the recent xarray update (reverting to 0.10.9 solves the problem).

The error is the famous (and very informative):

[Errno -101] NetCDF: HDF error:

I have not been able to create a MWE yet, but it has something to do with read -> close -> append workflows on netcdf4 files (the error happens at the "append" step). Possibly multiprocessing also plays a role, but I can't be sure yet.

I will try to find a way to reproduce this with a simple example, but this might take a while...

This is just a test to see if I can reproduce an HDF error reported in https://github.com/pydata/xarray/issues/2551

• [x] Closes https://github.com/pydata/xarray/issues/2497

The last time stickler had a look at our PRs is 8 days ago (https://github.com/pydata/xarray/pull/2415) : https://stickler-ci.com/repositories/26661-pydata-xarray

It looks like their bot is broken: https://github.com/stickler-ci

This is not very trustworthy - we could consider switching to https://pep8speaks.com/ maybe

I have a dataset with decreasing latitude coordinates.

With xarray v0.10.8, this is what happens when plotting:

But on latest master the image is now upside down:

Sorry if I missed a change along the way, I was off for a long time.

• [x] Closes https://github.com/pydata/xarray/issues/2422
• [x] Tests added
• [x] Tests passed
• [x] Fully documented, including whats-new.rst

https://github.com/pydata/xarray/pull/2294 introduced the new xincrease and yincrease kwargs and a behavior documented as:

xincrease : None, True, or False, optional Should the values on the x axes be increasing from left to right? if None, use the default for the matplotlib function.

The default was None, I suggest to set it to True: it was the default before https://github.com/pydata/xarray/pull/2294 and I think it is what most users expect.

See https://github.com/pydata/xarray/issues/2316

See https://github.com/pydata/xarray/issues/2306

This makes the builds a little bit faster. But we are still very close to the 900 s limit.

I tried to pin more packages this morning but this didn't work out. I'll merge this first (the latest build worked with this config, https://readthedocs.org/projects/xray/builds/) and try again at a later stage.

We are reaching the 900s build time limit on readthedocs more often than not (https://readthedocs.org/projects/xray/builds/). I have the same problem with all of my OS projects.

The bottleneck is the conda environment installation, which took 457s on the latest failed build. I'm going to try to spare some time in a subsequent PR, but we might have to get in touch with the RTD people to get more build time.

Simply because it makes more sense after all.

The link address won't change so it's not a biggie (http://xarray.pydata.org/en/latest/auto_gallery/index.html)

• [x] Closes https://github.com/pydata/xarray/issues/2250
• [x] Tests added
• [x] Tests passed
• [x] Fully documented, including whats-new.rst for all changes and api.rst for new API

Don't know if we want this, but it's possible to replace the default RTD icon. I'm not even sure if the cool kids are using bookmarks anymore, but I still do. This is how it looks like on my Firefox fav tab:

I'm seeing them from time to time on other repositories. One that seems reasonable and not toooo intrusive is stickler, for code style review: https://stickler-ci.com/

• [x] Closes https://github.com/pydata/xarray/issues/2088

https://github.com/pydata/xarray/commit/6402391cf206fd04c12d44773fecd9b42ea0c246 overwrote an array which was needed later on. In general the whole page was a bit messy with the same initial array being forgotten and then re-used much later.

This PR attempts to uniformize the variable names in order to make self-consistent subsection examples. I hope I didn't brake something on the way, but it looks ok on my local build.

(also renamed a gallery example plot which wasn't showing up correctly)

• [x] Closes https://github.com/pydata/xarray/issues/2002
• [x] Tests added
• [x] Tests passed
• [x] Fully documented, including whats-new.rst for all changes and api.rst for new API

Not sure yet if this is the best way to do this.

Code Sample, a copy-pastable example if possible

```python import xarray as xr import numpy as np

da = xr.DataArray(np.arange(9).reshape((1, 3, 3)))

da.plot() # works da.plot.imshow() # fails ``` Error log:

See https://github.com/pydata/xarray/pull/1957

• [x] Closes https://github.com/pydata/xarray/issues/1896
• [ ] Tests added
• [ ] Tests passed
• [x] Fully documented, including whats-new.rst for all changes and api.rst for new API

Instead of relying on seaborn per default we now use mpl and fall back to seaborn if cmap is not recognized (e.g. with cmap="husl").

Code Sample, a copy-pastable example if possible

```python import os import urllib.request import xarray as xr import matplotlib.pyplot as plt

Download the file from rasterio's repository

url = 'https://github.com/mapbox/rasterio/raw/master/tests/data/RGB.byte.tif' urllib.request.urlretrieve(url, 'RGB.byte.tif')

Read the data

da = xr.open_rasterio('RGB.byte.tif')

f, (ax1, ax2) = plt.subplots(1, 2, figsize=(9, 4)) da.plot.imshow(ax=ax1) (da / 255).plot.imshow(ax=ax2) plt.tight_layout() plt.show()

Delete the file

os.remove('RGB.byte.tif') ```

Problem description

In https://github.com/pydata/xarray/pull/1796, @Zac-HD added support for RGBA images. If an alpha channel is not found, it is added (code)

The problem is that adding this alpha channel requires the images to be normalized to 0-1, while plotting an image in 0-255 range without alpha channel works fine in matplotlib. Removing https://github.com/pydata/xarray/blob/master/xarray/plot/plot.py#L708-L715 would solve the problem, but I guess it was added for a reason.

@Zac-HD , thoughts?

The recent distributed update (1.20.2) broke a test: https://github.com/pydata/xarray/blob/master/xarray/tests/test_distributed.py#L57-L84

It fails with:

```

  assert s.task_state

E AttributeError: 'Scheduler' object has no attribute 'task_state' ```

Once https://github.com/pydata/xarray/pull/1817 and https://github.com/pydata/xarray/pull/1712 are merged it might be a good idea to revisit the tests to remove boilerplate code and try to generalize them.

• [x] Closes https://github.com/pydata/xarray/issues/1843
• [x] Tests added
• [x] Tests passed
• [x] Fully documented, including whats-new.rst for all changes and api.rst for new API

This PR restores the tests that were incorrectly removed in https://github.com/pydata/xarray/pull/1817 and adds the what's new entry I forgot in https://github.com/pydata/xarray/pull/1712

I think that pip install is the recommended way to install packages in dev mode (see also http://www.python3statement.org/practicalities/ )

https://github.com/pydata/xarray/pull/1796 introduced a bug in the doc gallery. This PR reverts the code to the previous greyscale example and adds a new case using imshow (the use case is different, as I tried to explain in the descriptions).

I also took care of https://github.com/pydata/xarray/issues/1789#issuecomment-356068358 : the docs should now build even when rasterio is not installed.

cc @Zac-HD , @shoyer

• [x] Closes https://github.com/pydata/xarray/issues/1686
• [x] Tests added / passed
• [x] Passes git diff upstream/master **/*py | flake8 --diff
• [ ] Fully documented, including whats-new.rst for all changes and api.rst for new API

Extends https://github.com/pydata/xarray/pull/1804 by adding some usage examples to the gallery.

Here is how the plot looks like:

cc @dcherian and @shoyer

Follow-up to https://github.com/pydata/xarray/pull/1049

Example:

```python import xarray as xr import numpy as np

ds = xr.Dataset() for vn in ['a', 'b', 'c']: ds[vn] = xr.DataArray(np.arange(10), dims=['t']) dsg = ds.groupby('t').mean() print(list(ds.variables.keys())) out : ['t', 'a', 'b', 'c'] print(list(dsg.variables.keys())) out: ['a', 'b', 'c', 't'] ```

Is it intentional? I think it is rather undesirable, but maybe there is some reason for this.

Code Sample, a copy-pastable example if possible

```python import numpy as np import xarray as xr

a = np.random.randn(4, 3) a[1, 1] = np.NaN da = xr.DataArray(a, dims=('y', 'x'), coords={'y':np.arange(4), 'x':np.arange(3)}) da.to_netcdf('test.nc')

with xr.open_dataarray('test.nc') as da: da.dropna(dim='x', how='any')

ValueError Traceback (most recent call last) <ipython-input-37-8d137cf3a813> in <module>() 8 9 with xr.open_dataarray('test.nc') as da: ---> 10 da.dropna(dim='x', how='any')

~/.pyvirtualenvs/py3/lib/python3.5/site-packages/xarray/core/dataarray.py in dropna(self, dim, how, thresh) 1158 DataArray 1159 """ -> 1160 ds = self._to_temp_dataset().dropna(dim, how=how, thresh=thresh) 1161 return self._from_temp_dataset(ds) 1162

~/.pyvirtualenvs/py3/lib/python3.5/site-packages/xarray/core/dataset.py in dropna(self, dim, how, thresh, subset) 2292 raise TypeError('must specify how or thresh') 2293 -> 2294 return self.isel(**{dim: mask}) 2295 2296 def fillna(self, value):

~/.pyvirtualenvs/py3/lib/python3.5/site-packages/xarray/core/dataset.py in isel(self, drop, **indexers) 1291 coord_names = set(variables).intersection(self._coord_names) 1292 selected = self._replace_vars_and_dims(variables, -> 1293 coord_names=coord_names) 1294 1295 # Extract coordinates from indexers

~/.pyvirtualenvs/py3/lib/python3.5/site-packages/xarray/core/dataset.py in _replace_vars_and_dims(self, variables, coord_names, dims, attrs, inplace) 598 """ 599 if dims is None: --> 600 dims = calculate_dimensions(variables) 601 if inplace: 602 self._dims = dims

~/.pyvirtualenvs/py3/lib/python3.5/site-packages/xarray/core/dataset.py in calculate_dimensions(variables) 111 raise ValueError('conflicting sizes for dimension %r: ' 112 'length %s on %r and length %s on %r' % --> 113 (dim, size, k, dims[dim], last_used[dim])) 114 return dims 115

ValueError: conflicting sizes for dimension 'y': length 2 on <this-array> and length 4 on 'y' ```

Problem description

See above. Note that the code runs when: - data is previously read into memory with load() - the DataArray is stored without coordinates (this is strange) - dropna is applied to 'y' instead of 'x'

Expected Output

This used to work in v0.9.6

Output of xr.show_versions()

I think this is a regression in the current 0.10.0rc1:

Code Sample

```python import xarray as xr ds = xr.open_dataset('cesm_data.nc', decode_cf=False) ds.temp.isel(time=ds.time < 274383) # throws an error

NotImplementedError Traceback (most recent call last) <ipython-input-18-a5c4179cd02d> in <module>() ----> 1 ds.temp.isel(time=ds.time < 274383)

~/.pyvirtualenvs/py3/lib/python3.5/site-packages/xarray/core/dataarray.py in isel(self, drop, indexers) 717 DataArray.sel 718 """ --> 719 ds = self._to_temp_dataset().isel(drop=drop, indexers) 720 return self._from_temp_dataset(ds) 721

~/.pyvirtualenvs/py3/lib/python3.5/site-packages/xarray/core/dataset.py in isel(self, drop, indexers) 1278 for name, var in iteritems(self._variables): 1279 var_indexers = {k: v for k, v in indexers_list if k in var.dims} -> 1280 new_var = var.isel(var_indexers) 1281 if not (drop and name in var_indexers): 1282 variables[name] = new_var

~/.pyvirtualenvs/py3/lib/python3.5/site-packages/xarray/core/variable.py in isel(self, **indexers) 771 if dim in indexers: 772 key[i] = indexers[dim] --> 773 return self[tuple(key)] 774 775 def squeeze(self, dim=None):

~/.pyvirtualenvs/py3/lib/python3.5/site-packages/xarray/core/variable.py in getitem(self, key) 595 """ 596 dims, index_tuple, new_order = self._broadcast_indexes(key) --> 597 data = self._indexable_data[index_tuple] 598 if new_order: 599 data = np.moveaxis(data, range(len(new_order)), new_order)

~/.pyvirtualenvs/py3/lib/python3.5/site-packages/xarray/core/indexing.py in getitem(self, key) 414 415 def getitem(self, key): --> 416 return type(self)(_wrap_numpy_scalars(self.array[key])) 417 418 def setitem(self, key, value):

~/.pyvirtualenvs/py3/lib/python3.5/site-packages/xarray/core/indexing.py in getitem(self, key) 394 395 def getitem(self, key): --> 396 return type(self)(_wrap_numpy_scalars(self.array[key])) 397 398 def setitem(self, key, value):

~/.pyvirtualenvs/py3/lib/python3.5/site-packages/xarray/core/indexing.py in getitem(self, key) 361 362 def getitem(self, key): --> 363 return type(self)(self.array, self._updated_key(key)) 364 365 def setitem(self, key, value):

~/.pyvirtualenvs/py3/lib/python3.5/site-packages/xarray/core/indexing.py in _updated_key(self, new_key) 336 raise NotImplementedError( 337 'Vectorized indexing for {} is not implemented. Load your ' --> 338 'data first with .load() or .compute().'.format(type(self))) 339 new_key = iter(expanded_indexer(new_key, self.ndim)) 340 key = []

NotImplementedError: Vectorized indexing for <class 'xarray.core.indexing.LazilyIndexedArray'> is not implemented. Load your data first with .load() or .compute().

``` Here is the file: cesm_data.nc.zip

Expected Output

This used to work in v0.9

Output of xr.show_versions()

While working on https://github.com/pydata/xarray/pull/1260 I came upon this which looks like a bug in caching:

```python import numpy as np import xarray as xr from xarray.core import indexing

nx, ny = 8, 10 data = np.arange(nx*ny).reshape(ny, nx) cached = indexing.MemoryCachedArray(data)

data = xr.DataArray(data=data, dims=('y', 'x')) cached = xr.DataArray(data=cached, dims=('y', 'x'))

a = data.isel(x=[2, 4], y=[3, 5]) b = cached.isel(x=[2, 4], y=[3, 5]) ``` The last line raises:

```

AssertionError Traceback (most recent call last) <ipython-input-13-45cd1493cf6b> in <module>() 11 12 a = data.isel(x=[2, 4], y=[3, 5]) ---> 13 b = cached.isel(x=[2, 4], y=[3, 5])

/home/mowglie/Documents/git/xarray/xarray/core/dataarray.py in isel(self, drop, indexers) 668 DataArray.sel 669 """ --> 670 ds = self._to_temp_dataset().isel(drop=drop, indexers) 671 return self._from_temp_dataset(ds) 672

/home/mowglie/Documents/git/xarray/xarray/core/dataset.py in isel(self, drop, indexers) 1141 for name, var in iteritems(self._variables): 1142 var_indexers = dict((k, v) for k, v in indexers if k in var.dims) -> 1143 new_var = var.isel(var_indexers) 1144 if not (drop and name in var_indexers): 1145 variables[name] = new_var

/home/mowglie/Documents/git/xarray/xarray/core/variable.py in isel(self, **indexers) 547 if dim in indexers: 548 key[i] = indexers[dim] --> 549 return self[tuple(key)] 550 551 def squeeze(self, dim=None):

/home/mowglie/Documents/git/xarray/xarray/core/variable.py in getitem(self, key) 380 # orthogonal indexing should ensure the dimensionality is consistent 381 if hasattr(values, 'ndim'): --> 382 assert values.ndim == len(dims), (values.ndim, len(dims)) 383 else: 384 assert len(dims) == 0, len(dims)

AssertionError: (1, 2) ```

The following code used to work well in v0.8.2:

```python import os import xarray as xr

path = 'test.nc' if os.path.exists(path): os.remove(path)

ds = xr.Dataset() ds['dim'] = ('dim', [0, 1, 2]) ds['var1'] = ('dim', [10, 11, 12]) ds.to_netcdf(path)

ds = xr.Dataset() ds['dim'] = ('dim', [0, 1, 2]) ds['var2'] = ('dim', [10, 11, 12]) ds.to_netcdf(path, 'a') ```

On master, it fails with:

```

RuntimeError Traceback (most recent call last) <ipython-input-1-fce5f5e876aa> in <module>() 14 ds['dim'] = ('dim', [0, 1, 2]) 15 ds['var2'] = ('dim', [10, 11, 12]) ---> 16 ds.to_netcdf(path, 'a')

/home/mowglie/Documents/git/xarray/xarray/core/dataset.py in to_netcdf(self, path, mode, format, group, engine, encoding) 927 from ..backends.api import to_netcdf 928 return to_netcdf(self, path, mode, format=format, group=group, --> 929 engine=engine, encoding=encoding) 930 931 def unicode(self):

/home/mowglie/Documents/git/xarray/xarray/backends/api.py in to_netcdf(dataset, path, mode, format, group, engine, writer, encoding) 563 store = store_cls(path, mode, format, group, writer) 564 try: --> 565 dataset.dump_to_store(store, sync=sync, encoding=encoding) 566 if isinstance(path, BytesIO): 567 return path.getvalue()

/home/mowglie/Documents/git/xarray/xarray/core/dataset.py in dump_to_store(self, store, encoder, sync, encoding) 873 variables, attrs = encoder(variables, attrs) 874 --> 875 store.store(variables, attrs, check_encoding) 876 if sync: 877 store.sync()

/home/mowglie/Documents/git/xarray/xarray/backends/common.py in store(self, variables, attributes, check_encoding_set) 219 cf_variables, cf_attrs = cf_encoder(variables, attributes) 220 AbstractWritableDataStore.store(self, cf_variables, cf_attrs, --> 221 check_encoding_set) 222 223

/home/mowglie/Documents/git/xarray/xarray/backends/common.py in store(self, variables, attributes, check_encoding_set) 194 def store(self, variables, attributes, check_encoding_set=frozenset()): 195 self.set_attributes(attributes) --> 196 self.set_variables(variables, check_encoding_set) 197 198 def set_attributes(self, attributes):

/home/mowglie/Documents/git/xarray/xarray/backends/common.py in set_variables(self, variables, check_encoding_set) 204 name = _encode_variable_name(vn) 205 check = vn in check_encoding_set --> 206 target, source = self.prepare_variable(name, v, check) 207 self.writer.add(source, target) 208

/home/mowglie/Documents/git/xarray/xarray/backends/netCDF4_.py in prepare_variable(self, name, variable, check_encoding) 293 endian='native', 294 least_significant_digit=encoding.get('least_significant_digit'), --> 295 fill_value=fill_value) 296 nc4_var.set_auto_maskandscale(False) 297

netCDF4/_netCDF4.pyx in netCDF4._netCDF4.Dataset.createVariable (netCDF4/_netCDF4.c:18740)()

netCDF4/_netCDF4.pyx in netCDF4._netCDF4.Variable.init (netCDF4/_netCDF4.c:30713)()

RuntimeError: NetCDF: String match to name in use ```

Some live news from the classroom. A student (who is learning python and xarray at the same time) wanted to compute the minimum of an array and forgot the parenthesis (quite a common mistake).

The printout in the notebook in that case is:

```python

ds.toa_sw_all_mon.min <bound method ImplementsArrayReduce._reduce_method.\<locals>.wrapped_func of \<xarray.DataArray 'toa_sw_all_mon' (month: 12, lat: 180, lon: 360)> [777600 values with dtype=float32] Coordinates: * month (month) float32 1.0 2.0 3.0 4.0 5.0 6.0 7.0 8.0 9.0 10.0 11.0 12.0 * lat (lat) float32 -89.5 -88.5 -87.5 -86.5 -85.5 -84.5 -83.5 -82.5 ... * lon (lon) float32 -179.5 -178.5 -177.5 -176.5 -175.5 -174.5 -173.5 ... Attributes: long_name: Top of The Atmosphere Shortwave Flux, Monthly Means, All-... standard_name: TOA Shortwave Flux - All-Sky CF_name: toa_outgoing_shortwave_flux IPCC_name: none units: W m-2 valid_min: 0.00000 valid_max: 600.000> ```

which, I had to agree, is hard to identify as being a function. It's a detail, but is it necessary to print the entire repr of the array/dataset in the function repr?

Addresses https://github.com/pydata/xarray/issues/752 and allows to pass kwargs to colorbar. For example, it is now possible to do:

``` python import numpy as np import matplotlib.pyplot as plt import xarray as xr

x, y = np.meshgrid(np.arange(12), np.arange(12)) z = xr.DataArray(np.sqrt(x2 + y2)) ds = z.to_dataset(name='z')

fig, ((ax1, ax2), (ax3, ax4)) = plt.subplots(2, 2, figsize=(12, 12)) ds.z.plot.contourf(ax=ax1) ds.z.plot.contourf(ax=ax2, cbar_kwargs={'orientation':'horizontal', 'label':'MyLabel'}) ds.z.plot.contourf(ax=ax3, cbar_ax=ax4, cbar_kwargs={'orientation':'horizontal', 'label':'Funny Cbar', 'drawedges':True}) plt.tight_layout() plt.show() ```

• [ ] Closes #xxxx
• [ ] Tests added / passed
• [ ] Passes git diff upstream/master | flake8 --diff
• [ ] Fully documented, including whats-new.rst for all changes and api.rst for new API

The pinned version of rasterio we use for the doc builds isn't available anymore: https://github.com/conda-forge/rasterio-feedstock/pull/36

Tests: - builds: https://readthedocs.org/projects/xray/builds/5796746/ - displays correctly: http://xarray.pydata.org/en/fix-docs/auto_gallery/plot_rasterio.html#sphx-glr-auto-gallery-plot-rasterio-py

This is needed for the doc recipe to run

Salem is likely to remain a small project but it has a more-than-one user base and might provide some inspiration for the future pangeo projects.

Follow-up to https://github.com/pydata/xarray/pull/1070

This is my first backend so this is going to be a bit more work than I expected, but with your help we should be able to get through this.

A long todo list: - [x] closes https://github.com/pydata/xarray/issues/790 - [x] add tests - [x] passes git diff upstream/master | flake8 --diff - [x] whatsnew entry - [x] wrap getitem for lazy data read - [x] ~~make coordinates variables lazy too~~

• [ ] Closes #xxxx
• [ ] Tests added / passed
• [x] Passes git diff upstream/master | flake8 --diff
• [x] Fully documented, including whats-new.rst for all changes and api.rst for new API

this is needed for the Rasterio PR (https://github.com/pydata/xarray/pull/1260#discussion_r118824181)

There was no test for these functionalities in the xarray test suite for Dataset either, so I just added a dummy test for one of the keywords.

• [x] Closes https://github.com/pydata/xarray/issues/1061
• [ ] Tests added / passed
• [ ] Passes git diff upstream/master | flake8 --diff
• [x] Fully documented, including whats-new.rst for all changes and api.rst for new API

This uses sphinx-gallery to illustrate some examples of the xarray workflow. See it rendered here

It's nicer for the gallery if there is a plot to draw in the end but it doesn't have to be a plot. Once the gallery gets more examples it's possible to sort them by topic (see the sphinx gallery documentation for an overview of the possibilities).

I could think of several other things we could add (including moving or copying @jhamman and @rabernat 's examples to the gallery), but I thought it's better to merge quickly in order to encourage the community to contribute with more examples before the next release.

[edit: this original PR was replaced with another change: "remove obsolete NetCDF4 Error catch"]

This allows to give a customized NetCDF4 object to the DataStore.

I needed this for Salem's diagnostic variables: https://github.com/fmaussion/salem/blob/master/salem/sio.py#L737-L751

My workaround (a very shallow subclass of NetCDF4DataStore) is fine for me, so I won't be offended if you decide not to merge.

On a related issue, the IndexError catch which happens here is also giving me trouble, since it hides bugs in my code.

I know I'm kind of misusing the backend, but I find it very useful the way it is. I'm open to any suggestion to make it more elegant. Thanks!

• [ ] Closes #xxxx
• [ ] Tests added / passed
• [ ] Passes git diff upstream/master | flake8 --diff
• [ ] Fully documented, including whats-new.rst for all changes and api.rst for new API

For some reason a plot from the examples was rendered statically instead of live. This is now fixed, together with an update of pandas which prevents reshaping operations on indexes.

Tested here

Not sure if this is an xarray issue, a netCDF4 or a h5py one, but I found that importing h5py is not a good idea if you want to write to netcdf4 afterwards:

python import h5py import xarray as xr ds = xr.Dataset({'x': [1, 2, 3]}) ds.to_netcdf('test.nc4')

Errors with:

```

RuntimeError Traceback (most recent call last) <ipython-input-7-924aff64174f> in <module>() 2 import xarray as xr 3 ds = xr.Dataset({'x': [1, 2, 3]}) ----> 4 ds.to_netcdf('test.nc4')

/home/mowglie/Documents/git/xarray/xarray/core/dataset.py in to_netcdf(self, path, mode, format, group, engine, encoding, unlimited_dims) 974 return to_netcdf(self, path, mode, format=format, group=group, 975 engine=engine, encoding=encoding, --> 976 unlimited_dims=unlimited_dims) 977 978 def unicode(self):

/home/mowglie/Documents/git/xarray/xarray/backends/api.py in to_netcdf(dataset, path_or_file, mode, format, group, engine, writer, encoding, unlimited_dims) 581 try: 582 dataset.dump_to_store(store, sync=sync, encoding=encoding, --> 583 unlimited_dims=unlimited_dims) 584 if path_or_file is None: 585 return target.getvalue()

/home/mowglie/Documents/git/xarray/xarray/core/dataset.py in dump_to_store(self, store, encoder, sync, encoding, unlimited_dims) 913 914 store.store(variables, attrs, check_encoding, --> 915 unlimited_dims=unlimited_dims) 916 if sync: 917 store.sync()

/home/mowglie/Documents/git/xarray/xarray/backends/common.py in store(self, variables, attributes, args, kwargs) 244 cf_variables, cf_attrs = cf_encoder(variables, attributes) 245 AbstractWritableDataStore.store(self, cf_variables, cf_attrs, --> 246 args, **kwargs) 247 248

/home/mowglie/Documents/git/xarray/xarray/backends/common.py in store(self, variables, attributes, check_encoding_set, unlimited_dims) 213 self.set_attributes(attributes) 214 self.set_variables(variables, check_encoding_set, --> 215 unlimited_dims=unlimited_dims) 216 217 def set_attributes(self, attributes):

/home/mowglie/Documents/git/xarray/xarray/backends/netCDF4_.py in set_variables(self, args, kwargs) 286 def set_variables(self, args, kwargs): 287 with self.ensure_open(autoclose=False): --> 288 super(NetCDF4DataStore, self).set_variables(*args, kwargs) 289 290 def prepare_variable(self, name, variable, check_encoding=False,

/usr/lib/python3.5/contextlib.py in exit(self, type, value, traceback) 75 value = type() 76 try: ---> 77 self.gen.throw(type, value, traceback) 78 raise RuntimeError("generator didn't stop after throw()") 79 except StopIteration as exc:

/home/mowglie/Documents/git/xarray/xarray/backends/common.py in ensure_open(self, autoclose) 282 self.close() 283 else: --> 284 yield 285 286 def assert_open(self):

/home/mowglie/Documents/git/xarray/xarray/backends/netCDF4_.py in set_variables(self, args, kwargs) 286 def set_variables(self, args, kwargs): 287 with self.ensure_open(autoclose=False): --> 288 super(NetCDF4DataStore, self).set_variables(*args, kwargs) 289 290 def prepare_variable(self, name, variable, check_encoding=False,

/home/mowglie/Documents/git/xarray/xarray/backends/common.py in set_variables(self, variables, check_encoding_set, unlimited_dims) 226 target, source = self.prepare_variable( 227 name, v, check, unlimited_dims=unlimited_dims) --> 228 self.writer.add(source, target) 229 230 def set_necessary_dimensions(self, variable, unlimited_dims=None):

/home/mowglie/Documents/git/xarray/xarray/backends/common.py in add(self, source, target) 167 else: 168 try: --> 169 target[...] = source 170 except TypeError: 171 # workaround for GH: scipy/scipy#6880

netCDF4/_netCDF4.pyx in netCDF4._netCDF4.Variable.setitem (netCDF4/_netCDF4.c:48315)()

netCDF4/_netCDF4.pyx in netCDF4._netCDF4.Variable._put (netCDF4/_netCDF4.c:49808)()

RuntimeError: NetCDF: HDF error ```

Note that using engine='scipy' or omitting to import h5py doesn't throw an error of course.

For the record: - h5py: 2.7.0 - xarray: 0.9.5-5-gfd6e36e - system: linux

• [ ] closes #xxxx
• [ ] tests added / passed
• [ ] passes git diff upstream/master | flake8 --diff
• [ ] whatsnew entry

Since the last pandas update I get:

/home/mowglie/Documents/git/xarray/xarray/core/formatting.py:16: FutureWarning: The pandas.tslib module is deprecated and will be removed in a future version. from pandas.tslib import OutOfBoundsDatetime

• [x] closes https://github.com/pydata/xarray/issues/1401
• [ ] tests added / passed
• [ ] passes git diff upstream/master | flake8 --diff
• [ ] whatsnew entry

I don't know if we need to install all packages to cover all the tests which rely on bottleneck, but this was the easiest.

This was introduced by https://github.com/pydata/xarray/pull/1368 (@shoyer)

Picking up a thread about the repr (https://github.com/pydata/xarray/issues/1199#issuecomment-272824929), I think it would be good to give a name to the two different types of coordinates in xarray.

Currently the doc says:

One dimensional coordinates with a name equal to their sole dimension (marked by * when printing a dataset or data array) take on a special meaning in xarray. They are used for label based indexing and alignment, like the index found on a pandas DataFrame or Series. Indeed, these “dimension” coordinates use a pandas.Index internally to store their values.

Other than for indexing, xarray does not make any direct use of the values associated with coordinates. Coordinates with names not matching a dimension are not used for alignment or indexing, nor are they required to match when doing arithmetic (see Coordinates).

The use of quotation marks in “dimension” coordinates makes the term imprecise. Should we simply call the former dimension coordinates and the latter optional coordinates?

This would also help to uniformize error reporting (e.g. https://github.com/pydata/xarray/pull/1291#discussion_r104261803)

• [x] closes https://github.com/pydata/xarray/issues/1295

This is a RTD problem so out of our control, but I'll leave this issue opened until it is resolved.

See RTD issue: https://github.com/rtfd/readthedocs.org/issues/2651

Sorry I missed that one when it was decided upon in https://github.com/pydata/xarray/pull/1017, but I think the changes in repr should be documented somewhere (at the minimum in the "Breaking Changes" section of what's new).

I just updated Salem for it to work well with xarray 0.9.0. The changes I had to make where quite small (that's a good thing), but it took me a bit of time to understand what was going on.

What I found confusing is following:

```python In [1]: import xarray as xr In [2]: ds = xr.DataArray([1, 2, 3]).to_dataset(name='var') In [3]: ds Out[3]: <xarray.Dataset> Dimensions: (dim_0: 3) Coordinates: o dim_0 (dim_0) - Data variables: var (dim_0) int64 1 2 3

In [4]: 'dim_0' in ds.coords Out[4]: False `dim_0is listed as coordinate, but'dim_0' in ds.coordsisFalse``. I think it should remain like this, but maybe we should document somewhere what the "o" and "*" mean?

(possibly here)

• [x] closes https://github.com/pydata/xarray/issues/1208
• [ ] tests added / passed
• [x] passes git diff upstream/master | flake8 --diff
• [ ] whatsnew entry

For some reason I can't reach the webpage but @shoyer seems to be able to do so: https://github.com/pydata/xarray/issues/1167#issuecomment-274966778

• [x] closes https://github.com/pydata/xarray/issues/1218
• [x] tests added / passed
• [x] whatsnew entry

This PR fixes a problem we had with ipython on RTD: https://github.com/ipython/ipython/issues/8733

We this simple workaround we can now use the latest ipython version. Thanks to @takluyver for his help!

This is a minor issue (sorry to be so picky about the repr ;) )

Small 2D coordinates are represented in a weird way:

```python In [1]: import xarray as xr

In [2]: a = np.array([[1.1, 2.2, 3.3], [4.4, 5.5, 6.6]])

In [3]: da = xr.DataArray(a, dims=['y', 'x'], coords={'xy':(['y', 'x'], a)})

In [4]: da Out[4]: <xarray.DataArray (y: 2, x: 3)> array([[ 1.1, 2.2, 3.3], [ 4.4, 5.5, 6.6]]) Coordinates: xy (y, x) float64 1.1 2.2 3.3 4.4 5.5 6.6 o y (y) - o x (x) - ```

This line here:

Coordinates: xy (y, x) float64 1.1 2.2 3.3 4.4 5.5 6.6 Is a bit confusing as it flattens the coordinates. It's not a big deal though, just something to keep in mind maybe.

follow-up to https://github.com/pydata/xarray/pull/1203

Just to see what happens with the upcoming matplotlib version 2

Follow-up to https://github.com/pydata/xarray/pull/1171#issuecomment-269556898

From the new docstring:

levels : int or list-like object, optional Split the colormap (cmap) into discrete color intervals. If an integer is provided, "nice" levels are chosen based on the data range: this can imply that the final number of levels is not exactly the expected one. Setting ``vmin`` and/or ``vmax`` with ``levels=N`` is equivalent to setting ``levels=np.linspace(vmin, vmax, N)``.

The logic overhead is quite simple, which is an argument in favor of this simple solution. It is consistent with mpl, as long as neither vmin or vmax are set.

It might change the outcome of some existing plots, though. Should I mention this in the "Breaking changes" section too?

This would help new users like https://github.com/pydata/xarray/issues/1113 and simplify the RTD build process (https://github.com/pydata/xarray/issues/1106).

The problem is that it is not as trivial as expected. On the latest master:

python import xarray as xr ds = xr.tutorial.load_dataset('rasm') ds.to_netcdf('rasm.nc', format='NETCDF3_CLASSIC', engine='scipy') Throws an error:

```python

ValueError Traceback (most recent call last) /home/mowglie/Documents/git/xarray/xarray/backends/api.py in to_netcdf(dataset, path, mode, format, group, engine, writer, encoding) 516 try: --> 517 dataset.dump_to_store(store, sync=sync, encoding=encoding) 518 if isinstance(path, BytesIO):

/home/mowglie/Documents/git/xarray/xarray/core/dataset.py in dump_to_store(self, store, encoder, sync, encoding) 754 if sync: --> 755 store.sync() 756

/home/mowglie/Documents/git/xarray/xarray/backends/scipy_.py in sync(self) 149 super(ScipyDataStore, self).sync() --> 150 self.ds.flush() 151

/home/mowglie/.pyvirtualenvs/py3/lib/python3.4/site-packages/scipy/io/netcdf.py in flush(self) 388 if hasattr(self, 'mode') and self.mode in 'wa': --> 389 self._write() 390 sync = flush

/home/mowglie/.pyvirtualenvs/py3/lib/python3.4/site-packages/scipy/io/netcdf.py in _write(self) 400 self._write_gatt_array() --> 401 self._write_var_array() 402

/home/mowglie/.pyvirtualenvs/py3/lib/python3.4/site-packages/scipy/io/netcdf.py in _write_var_array(self) 448 for name in variables: --> 449 self._write_var_metadata(name) 450 # Now that we have the metadata, we know the vsize of

/home/mowglie/.pyvirtualenvs/py3/lib/python3.4/site-packages/scipy/io/netcdf.py in _write_var_metadata(self, name) 466 for dimname in var.dimensions: --> 467 dimid = self._dims.index(dimname) 468 self._pack_int(dimid)

ValueError: '2' is not in list ```

Partly addresses https://github.com/pydata/xarray/issues/768#issuecomment-187226020 and https://github.com/pydata/xarray/issues/1154, mostly by making the tone less defensive. I also added that we recommend the netCDF format for IO. (I think that the format is powerful enough for being useful in other disciplines than geosciences).

I agree with @rabernat that the current page title (Serialization and IO) could be changed to something more accessible, but I don't know how. Two possibilities, both quite long (it looks a bit ugly on RTD): - "Reading and writing xarray data structures on disk" - "Reading/writing xarray data structures"

Sometimes it could be useful to have a view of all variables attributes at a glance. For example, this is the repr for ERA-Interim energy fluxes data:

(...) Data variables: slhf (month, latitude, longitude) float64 0.02852 0.02852 0.02852 ... tsr (month, latitude, longitude) float64 -0.0001912 -0.0001912 ... strd (month, latitude, longitude) float64 166.6 166.6 166.6 166.6 ... strc (month, latitude, longitude) float64 -66.23 -66.23 -66.23 ... tisr (month, latitude, longitude) float64 0.0 0.0 0.0 0.0 0.0 0.0 ... ssrd (month, latitude, longitude) float64 -0.0003951 -0.0003951 ... ssrc (month, latitude, longitude) float64 0.0 0.0 0.0 0.0 0.0 0.0 ... str (month, latitude, longitude) float64 -40.65 -40.65 -40.65 ... ttr (month, latitude, longitude) float64 -171.5 -171.5 -171.5 ... tsrc (month, latitude, longitude) float64 0.0 0.0 0.0 0.0 0.0 0.0 ... sshf (month, latitude, longitude) float64 10.46 10.46 10.46 10.46 ... ssr (month, latitude, longitude) float64 0.0 0.0 0.0 0.0 0.0 0.0 ... ttrc (month, latitude, longitude) float64 -174.9 -174.9 -174.9 ... (...) This is what my students will see when they explore the dataset for the first time. It could be nice to have a utility function (e.g. dumph or something) which would have a style closer to ncdump -h:

(...) double ttr(month, latitude, longitude) ; ttr:least_significant_digit = 2L ; ttr:units = "J m**-2" ; ttr:long_name = "Top net thermal radiation" ; ttr:standard_name = "toa_outgoing_longwave_flux" ; double tsrc(month, latitude, longitude) ; tsrc:least_significant_digit = 2L ; tsrc:units = "J m**-2" ; tsrc:long_name = "Top net solar radiation, clear sky" ; double sshf(month, latitude, longitude) ; sshf:least_significant_digit = 2L ; sshf:units = "J m**-2" ; sshf:long_name = "Surface sensible heat flux" ; sshf:standard_name = "surface_upward_sensible_heat_flux" ; double ssr(month, latitude, longitude) ; ssr:least_significant_digit = 2L ; ssr:units = "J m**-2" ; ssr:long_name = "Surface net solar radiation" ; ssr:standard_name = "surface_net_downward_shortwave_flux" ; double ttrc(month, latitude, longitude) ; ttrc:least_significant_digit = 2L ; ttrc:units = "J m**-2" ; ttrc:long_name = "Top net thermal radiation, clear sky" ; (...)

Or is there something like this already?

Extends and finishes https://github.com/pydata/xarray/pull/637

I chose to keep seaborn's convention for two reasons: - it doesn't break existing code - now that figsize is also available, I expect the size and aspect kwargs to be less used in non-facetgrid plots

Closes https://github.com/pydata/xarray/issues/897

fixes https://github.com/pydata/xarray/issues/1159

This implements a more robust solution based on the location of conf.py...

I've had a hard time finding out what was going on with missing RTD plots since we updated the packages until I came across this: https://github.com/ipython/ipython/issues/8733

The proposed fix creates the @savefig directory if not there. It works locally but I'm not sure if it will work on RTD (they might use a custom build directory).

The other changes are minor tweaks to remove some build warnings and/or making use of the latest sphinx Makefile. I also removed a part of the conf.py which I think is not needed anymore.

@shoyer do you remember what the inspect.findsource monkeypatch was good for?

Another attempt to solve a cryptic error: https://readthedocs.org/projects/xray/builds/4727274/

But this time we are getting closer! I've found a report on ipython (https://github.com/ipython/ipython/issues/8850), an I've been able to reproduce it locally and solve it by updating to ipython 5

sorry about the many PRs :-(

Let's see what happens with this one

Pinning python to 3.5 introduced new problems.

Despite the default python being 3 here and also in our RTD preferences (dashboard->advanced), RTD seems to still use py2 per default...

Anyway, going back to py2 is the easiest for now, I might come back to this when I have more time.

The idea came from here: https://github.com/Unidata/netcdf4-python/issues/574#issuecomment-235435628

I could test it on another repo, so I'm confident that this should work here too. Obviously, this is a temporary solution.

hopefully fixes https://github.com/pydata/xarray/issues/1106

Now that we seem to have everything working on RTD, re-pin package versions

(follow-up to https://github.com/pydata/xarray/pull/1101)

http://xarray.pydata.org/en/latest/whats-new.html See the examples at the bottom.

These are all old examples relying on the "xray" package. We can either remove these examples (my suggestion) or update them to use array.

As a general rule, I think that the what's new page shouldn't contain any code, at least not code that has to be run by RTD

There are some uses of xray that needed to be fixed.

WIP: merge after you decide what to do with https://github.com/pydata/xarray/issues/1135

Also, I'd be in favor of removing the two notebooks in https://github.com/pydata/xarray/tree/master/examples since they are also available in the docs (http://xarray.pydata.org/en/latest/examples.html). The example on RTD have the disadvantage that it's less easy to do copy/paste with them, but notebooks are more error prone since they need to be maintained externally.

This is to ping @ocefpaf on whether you have an idea on what's going on with netCDF4 on Read The Docs. See the import error here.

This is our conda config file: https://github.com/pydata/xarray/blob/master/doc/environment.yml

I've found related discussions here or here. Note that I never saw this when building salem's doc, which installs many more packages from conda-forge (see conf file).

Thanks a lot for your help! And no hurry.

```python import xarray as xr import netCDF4

print(xr.version) # 0.8.2-50-g57facab print(netCDF4.version) # 1.2.4

ds = xr.tutorial.load_dataset('rasm') /home/mowglie/Documents/git/xarray/xarray/conventions.py:389: RuntimeWarning: Unable to decode time axis into full numpy.datetime64 objects, continuing using dummy netCDF4.datetime objects instead, reason: dates out of range result = decode_cf_datetime(example_value, units, calendar) ```

I'll have a closer look.

Fixes https://github.com/pydata/xarray/issues/1111

As discussed here: https://github.com/pydata/xarray/issues/1106

I also added a section somewhere to print the packages versions on RTD. It will be useful to pin the versions again once it works, and is useful enough to stay on the docs I think. I just didn't really knew where to put it, other ideas welcome!

As discussed in https://github.com/pydata/xarray/pull/1079

I just left out a couple of nearly empty __init__s

I also removed a couple of unused imports

Addresses https://github.com/pydata/xarray/issues/781

@jhamman what do you think? I'm not sure if infer_interval_breaks is the best name for the keyword but I didn't come up with anything better than that. (maybe infer_coord_intervals?)

Sorry @shoyer you might get a few of these debugging PRs today again.

The multidimensional coords example is now built live. I tested this locally and tried to keep the look of the figures as close as possible to the original ones.

CC @rabernat @shoyer

(I also added a link to salem in the list of packages extending xarray - shameless self-promotion)

as discussed in https://github.com/pydata/xarray/pull/1098

Can someone reproduce this or is it just me?

```python import matplotlib.pyplot as plt import numpy as np import xarray as xr import cartopy.crs as ccrs

nlats, nlons = (241, 480) lats = np.linspace(90, -90, nlats, dtype=np.float32) lons = np.linspace(-180, 180-0.75, nlons, dtype=np.float32) l1, l2 = np.meshgrid(lons, lats) data = xr.DataArray(l1 + l2, [('latitude', lats), ('longitude', lons)])

f = plt.figure()

ax1 = plt.subplot(2, 1, 1, projection=ccrs.Robinson()) data.plot.contourf(ax=ax1, transform=ccrs.PlateCarree()); ax1.coastlines(color='grey'); ax1.gridlines();

data += 180 # this is the line causing the problem

ax2 = plt.subplot(2, 1, 2, projection=ccrs.Robinson()) data.plot.contourf(ax=ax2, transform=ccrs.PlateCarree()); ax2.coastlines(color='grey'); ax2.gridlines();

plt.show() ``` Gives:

Alternative (and much easier) implementation to https://github.com/pydata/xarray/pull/1048

The problem with the coordinate variables remains the same

addresses https://github.com/pydata/xarray/issues/1042

The problem is that sets aren't ordered. I tried something based on an OrderedSet recipe, but I'm not sure that this is the most elegant way to go.

A remaining issue is that the coord variable is still not in place (['t', 'a', 'b', 'c'] becomes ['a', 'b', 'c', 't']), but I'm not sure it's possible to do it better (it happens here).

Any idea welcome