so update from my side: i updated the package and now only get a df with unique_id, ds and y. not even the column for the index of the reconc method

but i dont get an error. it just finishes super fast and returns a df without my reconciled forecast values

ok, found the error myself. but i think a meaningful error message would be awesome here! the forecast column name of y_fitted and y_hat was not aligned. i called one fitted_values and the other fc_values or so. but they have to have the exact same name. this is easy if you just trained a model, but i am also using target values on higher levels to adapt my forecast from lower levels and therefore i had some mismatch in column names.

so i tried from memory to re-construct the example, but now using some example code from your tutorial page and changing the column names i get a very helpful error message. it is a bit weird, because i am pretty sure it was a mismtach of column names which created the issue for me. unfortunately i dont have much more time to look into this, but this is the log i am getting and it is pretty good and helpful:

ColumnNotFoundError: The following columns were not found: ['fitted']

Hint: Did you mean one of these columns: ['unique_id', 'ds', 'y', 'forecast']?