merge two datasets gives unexpected dimensions

[shuai-zhou]

I have two datasets, one is ds_prcp (precipitation data), another is ds_tmean (mean temperature data), both have exactly the same dimensions (longitude: 1405, latitude: 621, t: 12)

I want to combine the two datasets using the merge method: ds = xr.merge([ds_prcp, ds_tmean]), but it gives the wrong dimensions:

The dimensions should be still (longitude: 1405, latitude: 621, t: 12). What is the right way to do this? Thanks.

[keewis]

just a quick guess: maybe your coordinates are not exactly equal? In other words, I think this is a floating-point issue.

To check that, try ds_prcp.longitude.equals(ds_tmean.longitude), and you can also try comparing the output of ds_prcp.longitude.item(0) with ds_tmean.longitude.item(0).

If you are sure that the coordinates should be exactly equal, you can override the coordinates of one dataset, e.g.:

ds_tmean = ds_tmean.assign_coords(longitude=ds_prcp.longitude, latitude=ds_prcp.latitude)

[dcherian]

Passing join="override" to merge will do what keewis suggested.

[shuai-zhou]

Thanks to both suggestions. Passing join="override" does the work! Thanks.