Fill missing values in metric arrays? [FEATURE]

[jbusecke]

As mentioned in #182 here xgcm currently fails out when the metrics we use have missing values.

We could very easily fix this by adding a .fillna(0) here but I am not sure that we want to do that every time. How much performance will this cost?

The other option I see is to catch the exception and provide a more helpful error message specific to xgcm? This could sound somthing like this:

xgcm cannot use the metric x, because it has missing values. Please fill missing values with ds['x'].fillna(0) before creating the grid.

Thoughts?

[raphaeldussin]

@jbusecke this may be a little out-of-scope but one issue with filling missing values in coordinates with zeros is that it can mess up plots quite a bit. I wonder if using some type of extrapolation for coordinates (e.g. https://github.com/ajdawson/gridfill or https://github.com/raphaeldussin/HCtFlood) would be better?

[jbusecke]

Good point. In that case, I think I am getting more and more convinced to not alter the data automatically but instead build a better warning.

These packages are very interesting for the stuff I am doing over at cmip6_preprocessing though! Thanks for the links.

[github-actions]

This issue has been marked 'stale' due to lack of recent activity. If there is no further activity, the issue will be closed in another 30 days. Thank you for your contribution!

[github-actions]

This issue has been closed due to inactivity. If you feel this is in error, please reopen the issue or file a new issue with the relevant details.