How to append to a Dask Dataframe

[KonGian]

Thank you very much for your help.
Assuming that the point with the duplicates you have mentioned is not a problem and that the results can fit in a pandas Dataframe but the input Dataframe should be a Dask Dataframe, is there anything that I could do to make the computation of duplicates faster?

@KonGian

combinations = list(itertools.combinations(ddf.columns, 2))
for combination in combinations:
    distinct_tuples = ddf.drop_duplicates(subset=combination).shape[0].compute()
    distincts = np.append(distincts, [distinct_tuples])

In the previous code snippet combinations and distincts can fit in memory, and ddf is a dask dataframe.
This loop runs very slow when running in dask distributed environment. Is there any way to make this for loop run in parallel using Dask?