TypeError: unhashable type: 'list' #542

fipoucat · 2023-09-14T09:21:04Z

fipoucat
Sep 14, 2023

Question criteria

I added a descriptive title here.
I searched the xCDAT GitHub Discussions to find a similar question and didn't find it.
I searched the xCDAT documentation and/or Xarray documentation.

Describe your question

I installed xcat and started testing the examples but noticed an error unhashable type.

ds = xc.open_dataset(
    filepath,
    add_bounds=True,
    decode_times=True,
    center_times=True
)

Traceback (most recent call last):

  Cell In [3], line 3
    ds = xc.open_dataset(

  File ~/anaconda3/lib/python3.10/site-packages/xcdat/dataset.py:126 in open_dataset
    ds = _postprocess_dataset(ds, data_var, center_times, add_bounds, lon_orient)

  File ~/anaconda3/lib/python3.10/site-packages/xcdat/dataset.py:611 in _postprocess_dataset
    ds = ds.bounds.add_missing_bounds(axes=add_bounds)

  File ~/anaconda3/lib/python3.10/site-packages/xcdat/bounds.py:159 in add_missing_bounds
    for axis in axes:

TypeError: 'bool' object is not iterable

Any hint?

Are there are any possible answers you came across?

No response

Minimal Complete Verifiable Example (MVCE)

No response

Relevant log output

No response

Environment

No response

Anything else we need to know?

No response

Answered by pochedls

Sep 18, 2023

@fipoucat – I think you are likely using version 0.6.0rc1 (this is a "release candidate," which means it is being tested; I think it will be released as v0.6.0 soon). I can reproduce your issue with the release candidate version, but not with v0.5.

You can: a) use version 0.5 for now, b) remove the add_bounds=True argument in the release candidate version, or c) wait until v0.6 is released (I think Tom is going to address the issue you raise here).

View full answer

jypeter · 2023-09-14T11:38:10Z

jypeter
Sep 14, 2023

@fipoucat Which version of xcdat are you using? The 0.5.0 version, or a more recent development version?

The release candidate discussion #517 mentions add_bounds default value changed from add_bounds=True to add_bounds=["X","Y"]. This indeed quite different from what is described in the current open_dataset documentation

You may want to remove add_bounds=True in your call to xc.open_dataset, or supply an explicit list of the axes where you want to add bounds

2 replies

tomvothecoder Sep 14, 2023
Maintainer

Thanks for the answer @jypeter, you are correct here. v0.5.0 updates add_bounds=True to add_bounds=["X","Y"], with True being dropped as a valid type.

This brings up a good point that we should have probably added a deprecation warning and make add_bounds=True equivalent to add_bounds=["X","Y"]. We'll be more careful about these breaking changes since the package is starting to mature.

In the meantime, I'll add a patch so that users who upgrade from an older version to v0.5.0 don't run into this issue. Just an FYI @xCDAT/core-developers.

tomvothecoder Sep 14, 2023
Maintainer

Actually, v0.5.0 doesn't break and still uses add_bounds=True

xcdat/xcdat/dataset.py

Line 43 in d0a4892

add_bounds: bool = True,

v0.6.0rc1 (a release candidate) introduces this change. We just need to update v0.6.0 to include a deprecation warning if users pass True still.

fipoucat · 2023-09-14T17:40:25Z

fipoucat
Sep 14, 2023
Author

Thank you all, well noted.

…

On Thu, Sep 14, 2023 at 4:36 PM Tom Vo ***@***.***> wrote: Thanks for the answer @jypeter <https://github.com/jypeter>, you are correct here. v0.5.0 updates add_bounds=True to add_bounds=["X","Y"], with True being dropped as a valid type. This brings up a good point that we should have probably added a deprecation warning and make add_bounds=True equivalent to add_bounds=["X","Y"]. We'll be more careful about these breaking changes since the package is starting to mature. *In the meantime, I'll add a patch so that users who upgrade from an older version to v0.5.0 don't run into this issue.* Just an FYI @xCDAT/core-developers <https://github.com/orgs/xCDAT/teams/core-developers>. — Reply to this email directly, view it on GitHub <#542 (reply in thread)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AAXY6LN6C6NMIYEV6HRZUBDX2MW7HANCNFSM6AAAAAA4XY5WKM> . You are receiving this because you were mentioned.Message ID: ***@***.***>

1 reply

pochedls Sep 17, 2023
Collaborator

@fipoucat – were you using v0.5.0 (this will help us to understand if this is due to new changes or some other issue)? You can print the version with:

import xcdat as xc
print(xc.__version__)

fipoucat · 2023-09-18T11:59:30Z

fipoucat
Sep 18, 2023
Author

Yes, I was using v0.5.0

…

On Sun, Sep 17, 2023 at 10:47 PM Stephen Po-Chedley < ***@***.***> wrote: @fipoucat <https://github.com/fipoucat> – were you using v0.5.0 (this will help us to understand if this is due to new changes or some other issue)? You can print the version with: import xcdat as xc print(xc.__version__) — Reply to this email directly, view it on GitHub <#542 (reply in thread)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AAXY6LOQKPYI5NIN7B3MGR3X254ZRANCNFSM6AAAAAA4XY5WKM> . You are receiving this because you were mentioned.Message ID: ***@***.***>

1 reply

pochedls Sep 18, 2023
Collaborator

In tracing this down, it seems like you were using v0.6.0rc1, because you get the TypeError on line 159 (for axis in axes) of bounds.py, which doesn't seem to match v0.5.0. I am a little confused about what is going on.

Is the dataset you were using publicly available? I'm curious if the issue you had can be reproduced.

...and...are you still having this issue? Or did changing the add_bounds argument address your issue (if so, please let us know how)?

fipoucat · 2023-09-18T14:42:15Z

fipoucat
Sep 18, 2023
Author

I am using the example in the user guide ```python import xcdat as xc import xesmf # We import these packages specifically for plotting. It is not required to use xcdat. import matplotlib.pyplot as plt import pandas as pd filepath = " https://esgf-data1.llnl.gov/thredds/dodsC/css03_data/CMIP6/CMIP/CSIRO/ACCESS-ESM1-5/historical/r10i1p1f1/Amon/tas/gn/v20200605/tas_Amon_ACCESS-ESM1-5_historical_r10i1p1f1_gn_185001-201412.nc " ```

…

On Mon, Sep 18, 2023 at 2:25 PM Stephen Po-Chedley ***@***.***> wrote: In tracing this down, it *seems* like you were using v0.6.0rc1, because you get the TypeError on line 159 (for axis in axes) <https://github.com/xCDAT/xcdat/blob/v0.6.0rc1/xcdat/bounds.py#L159> of bounds.py, which doesn't seem to match v0.5.0 <https://github.com/xCDAT/xcdat/blob/main/xcdat/bounds.py#L159>. I am a little confused about what is going on. Is the dataset you were using publicly available? I'm curious if the issue you had can be reproduced. ...and...are you still having this issue? Or did changing the add_bounds argument address your issue (if so, please let us know how)? — Reply to this email directly, view it on GitHub <#542 (reply in thread)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AAXY6LIEBVV5SKARY7VSPR3X3BKVPANCNFSM6AAAAAA4XY5WKM> . You are receiving this because you were mentioned.Message ID: ***@***.***>

2 replies

pochedls Sep 18, 2023
Collaborator

@fipoucat – I think you are likely using version 0.6.0rc1 (this is a "release candidate," which means it is being tested; I think it will be released as v0.6.0 soon). I can reproduce your issue with the release candidate version, but not with v0.5.

You can: a) use version 0.5 for now, b) remove the add_bounds=True argument in the release candidate version, or c) wait until v0.6 is released (I think Tom is going to address the issue you raise here).

Answer selected by tomvothecoder

tomvothecoder Sep 18, 2023
Maintainer

Thanks @pochedls. Yes I'm certain @fipoucat is using xcdat=0.6.0rc1. The installed version can be checked with this comment if we want to be sure.

b) remove the add_bounds=True argument in the release candidate version

Or replace add_bounds=True with add_bounds=["X", "Y"] if you want to add bounds if they are missing. "T" and "Z" are also valid axes.

c) wait until v0.6 is released (I think Tom is going to address the issue you raise here).

I merged #546 to address this issue, which will be included in v0.6.0. add_bounds=True and add_bounds=False will be deprecated in v0.7.0.

fipoucat · 2023-09-18T16:25:23Z

fipoucat
Sep 18, 2023
Author

Thank you, that sounds good.

…

On Mon, Sep 18, 2023 at 4:07 PM Tom Vo ***@***.***> wrote: Thanks @pochedls <https://github.com/pochedls>. Yes I'm certain @fipoucat <https://github.com/fipoucat> is using xcdat=0.6.0rc1. The installed version can be checked with this comment <#542 (reply in thread)> if we want to be sure. b) remove the add_bounds=True argument in the release candidate version Or replace add_bounds=True with add_bounds=["X", "Y"] if you want to add bounds. "T" and "Z" are also valid axes. c) wait until v0.6 is released (I think Tom is going to address the issue you raise here). I merged #546 <#546> to address this issue. add_bounds=True and add_bounds=False will be deprecated in v0.7.0. — Reply to this email directly, view it on GitHub <#542 (reply in thread)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AAXY6LKZ6R2PKJFHAKDKKOLX3BWSLANCNFSM6AAAAAA4XY5WKM> . You are receiving this because you were mentioned.Message ID: ***@***.***>

0 replies

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

TypeError: unhashable type: 'list' #542

{{title}}

{{editor}}'s edit

{{editor}}'s edit

Replies: 5 comments 6 replies

{{title}}

{{title}}

{{title}}

{{title}}

{{title}}

{{title}}

{{title}}

{{title}}

{{editor}}'s edit

{{editor}}'s edit

{{title}}

{{title}}

{{editor}}'s edit

{{editor}}'s edit

{{title}}

Select a reply

TypeError: unhashable type: 'list' #542

fipoucat Sep 14, 2023

Question criteria

Describe your question

Are there are any possible answers you came across?

Minimal Complete Verifiable Example (MVCE)

Relevant log output

Environment

Anything else we need to know?

Replies: 5 comments · 6 replies

jypeter Sep 14, 2023

tomvothecoder Sep 14, 2023 Maintainer

tomvothecoder Sep 14, 2023 Maintainer

fipoucat Sep 14, 2023 Author

pochedls Sep 17, 2023 Collaborator

fipoucat Sep 18, 2023 Author

pochedls Sep 18, 2023 Collaborator

fipoucat Sep 18, 2023 Author

pochedls Sep 18, 2023 Collaborator

tomvothecoder Sep 18, 2023 Maintainer

fipoucat Sep 18, 2023 Author

fipoucat
Sep 14, 2023

Replies: 5 comments 6 replies

jypeter
Sep 14, 2023

tomvothecoder Sep 14, 2023
Maintainer

tomvothecoder Sep 14, 2023
Maintainer

fipoucat
Sep 14, 2023
Author

pochedls Sep 17, 2023
Collaborator

fipoucat
Sep 18, 2023
Author

pochedls Sep 18, 2023
Collaborator

fipoucat
Sep 18, 2023
Author

pochedls Sep 18, 2023
Collaborator

tomvothecoder Sep 18, 2023
Maintainer

fipoucat
Sep 18, 2023
Author