spark scala column to counter of column unique values

Question

How can I correctly get column values as Map(k->v) where k is unique value and v is occurence count? I do it within groupby.

val getMapUDF = udf((arr: Arra