-
Notifications
You must be signed in to change notification settings - Fork 303
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
olci_l1b-reader and xarray=2023.12.0 are seemingly incompatible #2705
Comments
Could you try loading one of your files directly without satpy by doing: import xarray as xr
dataset = xr.open_dataset("/path/to/your/file.nc", decode_cf=True, mask_and_scale=True, chunks="auto") And see if that fails? As far as I can tell nothing in Satpy is referencing a "bands" dimension in this reader or before so this seems like something that changed in xarray that isn't compatible with the files. That would mean a fix in xarray is required. Can you also confirm what the latest (most recent) version of xarray is where this works? |
You are right, it also fails. It does work with It's I should have checked this before, feel free to close this here. |
Could you file a bug with xarray and link to it here? Could you also for the record paste a |
Yes, but I will have to do so tomorrow. That's the dump:
|
Oh yeah so xarray doesn't like this. Maybe this is a bad assumption on xarray's part that a variable can't have the same dimension twice. I could see how that breaks their model. I assume this is an accurate description from the OLCI file's point of view. @mraspaud this was your reader, right? |
@fwfichtner Were you able to file a bug with xarray? |
@djhoese nope, I noticed that they did quite some changes to handling dimensions in the recent update and that it takes me a bit longer to understand what their roadmap was that I was willing to spend during the holidays. Reference: pydata/xarray#8491 |
A happy new year @djhoese! I opened an issue upstream pydata/xarray#8579 and I am not sure how to continue from there, but I guess this has to be changed here accordingly. |
@fwfichtner Thanks for filing the issue. Nice that the xarray team acknowledges this as a problem. I see you have a workaround for the problem in that issue. Is this something we need to implement in the OLCI reader or is it ok for you to pin xarray and wait for the fix on their side? |
Yes, great @mraspaud, I was a little fast accepting our fate here. I currently pinned xarray and I am happy to wait until this is fixed assuming that this will happen within some foreseeable future. |
Any updates on this? Still seems to be broken today... |
Not that I know, we have to wait if anything happens in pydata/xarray#8579. We currently pinned xarray to |
Looks like it's not really prioritized on the xarray side, so if we want things to move, we probably have get our hands dirty and give a go at a pull request there... |
I don't have a lot of time to pick that up myself in the coming weeks, but would you be able to pinpoint what exactly needs to be done in xarray? I guess it is not just removing the raise. |
Any updates on this issue other than rolling back xarray? I am facing the same issue with Sentinel 3 OLCI |
@varisht-tathya we are still waiting on pydata/xarray#8579 to be solved... which won't happen unless someone takes the time to work on it... |
@fwfichtner @varisht-tathya just a heads up that there is now a pull request in xarray that fixes this: pydata/xarray#9099 |
@mraspaud awesome, thank you a lot! |
The PR in xarray has now been merged, so the next release should fix this problem. Hence closing this issue. |
Thanks again @mraspaud, now it has also been released |
Great, so this should now work with the conda version of xarray latest? Good to hear! |
Describe the bug
See below, this is resolved by pinning
xarray
to an earlier versionTo Reproduce
Expected behavior
Results in:
Actual results
Text output of actual results or error messages including full tracebacks if applicable.
Environment Info:
The text was updated successfully, but these errors were encountered: