We use the API to download netcdf files for datasets like reanalysis-era5-single-levels-monthly-means, and among the variables we request are k_index and total_totals_index. The subdataset names in the returned netcdf files are expected to be kx and totalx, respectively, based on table 2 here.
However, the file we get back instead contains the time-mean versions of those parameters, avg_kx and avg_totalx. Is this because the dataset is monthly means? We’ve only started validating our downloads by checking the subdataset names so it’s possible that we’ve only run into this now because we only counted the number of subdatasets before.