count distinct window function Databricks

Question

I am implementing count distinct window functions in Databricks.

select *,count(distinct Marks) over(partition by Name) from data

It seems that count distinct is not supported in Databricks, how can I replicate the same query in databricks.

Answer 1

Using collect_set + size functions:

select *, size(collect_set(Marks)) over(partition by Name) from data