xarray issue?

[fanmei]

I got the below error message after running one ACT example. This is new to me. I used to run this example without any problems. I have recreated an environment and reinstalled metpy and act. Still get the same error message. Any suggestions?

%run plot_wind_rose.py

---

ValueError Traceback (most recent call last)
File D:\Python_Practice\plot_wind_rose.py:16
12 import numpy as np
14 from matplotlib import pyplot as plt
---> 16 sonde_ds = act.io.armfiles.read_netcdf(
17 act.tests.sample_files.EXAMPLE_TWP_SONDE_WILDCARD)
19 WindDisplay = act.plotting.WindRoseDisplay(sonde_ds, figsize=(8, 10), subplot_shape=(2,))
20 WindDisplay.plot('deg', 'wspd',
21 spd_bins=np.linspace(0, 25, 5), num_dirs=30,
22 tick_interval=2, subplot_index=(0,))

File D:\Anac3\envs\aaf\Lib\site-packages\act\io\armfiles.py:141, in read_netcdf(filenames, concat_dim, return_None, combine, use_cftime, cftime_to_datetime64, **kwargs)
137 arm_ds = xr.open_mfdataset(filenames, **kwargs)
139 else:
140 # When all else fails raise the orginal exception
--> 141 raise exception
143 # Xarray has issues reading a CF formatted time units string if it contains
144 # timezone offset without a [+|-] preceeding timezone offset.
145 # pydata/xarray#3644
(...)
149 # does not preserve the time past ms precision. We will use ms precision for
150 # the conversion.
151 desired_time_precision = 'datetime64[ms]'

File D:\Anac3\envs\aaf\Lib\site-packages\act\io\armfiles.py:119, in read_netcdf(filenames, concat_dim, return_None, combine, use_cftime, cftime_to_datetime64, **kwargs)
115 except_tuple = except_tuple + (FileNotFoundError, OSError)
117 try:
118 # Read data file with Xarray function
--> 119 arm_ds = xr.open_mfdataset(filenames, **kwargs)
121 except except_tuple as exception:
122 # If requested return None for File not found error
123 if type(exception).name == 'FileNotFoundError':

File D:\Anac3\envs\aaf\Lib\site-packages\xarray\backends\api.py:976, in open_mfdataset(paths, chunks, concat_dim, compat, preprocess, engine, data_vars, coords, combine, parallel, join, attrs_file, combine_attrs, **kwargs)
971 ids, paths = (
972 list(combined_ids_paths.keys()),
973 list(combined_ids_paths.values()),
974 )
975 elif combine == "by_coords" and concat_dim is not None:
--> 976 raise ValueError(
977 "When combine='by_coords', passing a value for concat_dim has no "
978 "effect. To manually combine along a specific dimension you should "
979 "instead specify combine='nested' along with a value for concat_dim.",
980 )
982 open_kwargs = dict(engine=engine, chunks=chunks or {}, **kwargs)
984 if parallel:

ValueError: When combine='by_coords', passing a value for concat_dim has no effect. To manually combine along a specific dimension you should instead specify combine='nested' along with a value for concat_dim.

[AdamTheisen]

@fanmei can you let us know what version python and xarray you are running?

[fanmei]

python version : 3.9.13.final.0 xarray 2022.10.0 From: Adam Theisen ***@***.***> Sent: Tuesday, November 1, 2022 5:11 PM To: ARM-DOE/ACT ***@***.***> Cc: Mei, Fan ***@***.***>; Mention ***@***.***> Subject: Re: [ARM-DOE/ACT] xarray issue? (Discussion #546) Check twice before you click! This email originated from outside PNNL. @fanmei<https://gcc02.safelinks.protection.outlook.com/?url=https%3A%2F%2Fgithub.com%2Ffanmei&data=05%7C01%7Cfan.mei%40pnnl.gov%7C1b58d66986be41686ec108dabc2c160a%7Cd6faa5f90ae240338c0130048a38deeb%7C0%7C0%7C638029194784950640%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C3000%7C%7C%7C&sdata=IAU8azvcTmhPUoB5hsuAIZnat%2FH6LtlY4AserXTV8m8%3D&reserved=0> can you let us know what version python and xarray you are running? - Reply to this email directly, view it on GitHub<https://gcc02.safelinks.protection.outlook.com/?url=https%3A%2F%2Fgithub.com%2FARM-DOE%2FACT%2Fdiscussions%2F546%23discussioncomment-4029399&data=05%7C01%7Cfan.mei%40pnnl.gov%7C1b58d66986be41686ec108dabc2c160a%7Cd6faa5f90ae240338c0130048a38deeb%7C0%7C0%7C638029194784950640%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C3000%7C%7C%7C&sdata=rYU%2BOQ24g%2FjcDE%2B%2Fy%2F5%2FmhJcHglERA2NHQgaWEM280o%3D&reserved=0>, or unsubscribe<https://gcc02.safelinks.protection.outlook.com/?url=https%3A%2F%2Fgithub.com%2Fnotifications%2Funsubscribe-auth%2FAEAKS3KJ5D7IZISM5I32IULWGFFLHANCNFSM6AAAAAARUG24IA&data=05%7C01%7Cfan.mei%40pnnl.gov%7C1b58d66986be41686ec108dabc2c160a%7Cd6faa5f90ae240338c0130048a38deeb%7C0%7C0%7C638029194784950640%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C3000%7C%7C%7C&sdata=kBNyXhCvlS%2BuxwG8SaKs3GBA%2BjSfUXw%2FvG17v9Kv9u0%3D&reserved=0>. You are receiving this because you were mentioned.Message ID: ***@***.******@***.***>>

[AdamTheisen]

Pulling @mgrover1 and @zssherman into this discussion as well.

[mgrover1]

These seem like two separate issues.. @fanmei can you please copy and paste the full error from the first lines?

import act

sonde_ds = act.io.armfiles.read_netcdf(act.tests.sample_files.EXAMPLE_TWP_SONDE_WILDCARD)

Let's start there.

[fanmei]

It worked last week, but the whole jupyter lab was crashed. I don't know if it is related to the system update in my computer.
I have to remove/reinstall the package. Then, I start to get the error message.

I put your two lines in a notebook (jupyter lab).
here is the error message.

---

ValueError Traceback (most recent call last)
Cell In [1], line 3
1 import act
----> 3 sonde_ds = act.io.armfiles.read_netcdf(act.tests.sample_files.EXAMPLE_TWP_SONDE_WILDCARD)

File D:\Anac3\envs\aaf\Lib\site-packages\act\io\armfiles.py:141, in read_netcdf(filenames, concat_dim, return_None, combine, use_cftime, cftime_to_datetime64, **kwargs)
137 arm_ds = xr.open_mfdataset(filenames, **kwargs)
139 else:
140 # When all else fails raise the orginal exception
--> 141 raise exception
143 # Xarray has issues reading a CF formatted time units string if it contains
144 # timezone offset without a [+|-] preceeding timezone offset.
145 # pydata/xarray#3644
(...)
149 # does not preserve the time past ms precision. We will use ms precision for
150 # the conversion.
151 desired_time_precision = 'datetime64[ms]'

File D:\Anac3\envs\aaf\Lib\site-packages\act\io\armfiles.py:119, in read_netcdf(filenames, concat_dim, return_None, combine, use_cftime, cftime_to_datetime64, **kwargs)
115 except_tuple = except_tuple + (FileNotFoundError, OSError)
117 try:
118 # Read data file with Xarray function
--> 119 arm_ds = xr.open_mfdataset(filenames, **kwargs)
121 except except_tuple as exception:
122 # If requested return None for File not found error
123 if type(exception).name == 'FileNotFoundError':

File D:\Anac3\envs\aaf\Lib\site-packages\xarray\backends\api.py:976, in open_mfdataset(paths, chunks, concat_dim, compat, preprocess, engine, data_vars, coords, combine, parallel, join, attrs_file, combine_attrs, **kwargs)
971 ids, paths = (
972 list(combined_ids_paths.keys()),
973 list(combined_ids_paths.values()),
974 )
975 elif combine == "by_coords" and concat_dim is not None:
--> 976 raise ValueError(
977 "When combine='by_coords', passing a value for concat_dim has no "
978 "effect. To manually combine along a specific dimension you should "
979 "instead specify combine='nested' along with a value for concat_dim.",
980 )
982 open_kwargs = dict(engine=engine, chunks=chunks or {}, **kwargs)
984 if parallel:

ValueError: When combine='by_coords', passing a value for concat_dim has no effect. To manually combine along a specific dimension you should instead specify combine='nested' along with a value for concat_dim.

[AdamTheisen]

@fanmei I have not been able to recreate this. Do you have the cftime package installed?

conda install -c conda-forge cftime

[fanmei]

which conda returns "which : The term 'which' is not recognized as the name of a cmdlet, "

[mgrover1]

Which computer are you running on? Mac, Linux, Windows?

[fanmei]

Windows

[mgrover1]

The last thing I would try is going back to Xarray version 2022.09.0. Other than that, unfortunately, I cannot reproduce your issue.

[fanmei]

It finally worked!!! thank you so much, Max

I guess that it is really sensitive to the versions.

[AdamTheisen]

@fanmei did you ever get ACT updated to the latest version? There's been a lot added since v0.2!

conda install act-atmos=1.2.0

You might need to do the following since it's already installed.
conda install act-atmos=1.2.0 --force

[fanmei]

Yes, after updating act to 1.2.0 and change the python and xarray version. All scripts work great!

[fanmei]

Thank you so much for all your help and support, Adam and Max