The difference between countDistinct and distinct.count

Question

Why do I get different outputs for ..agg(countDistinct(\"member_id\") as \"count\") and ..distinct.count? Is the difference the same as between

Answer 1

df.agg(countDistinct("member_id") as "count")

returns the number of distinct values of the member_id column, ignoring all other columns, while

df.distinct.count

will count the number of distinct records in the DataFrame - where "distinct" means identical in values of all columns.

So, for example, the DataFrame:

+-----------+---------+
|member_name|member_id|
+-----------+---------+
|          a|        1|
|          b|        1|
|          b|        1|
+-----------+---------+

has only one distinct member_id value but two distinct records, so the agg option would return 1 while the latter would return 2.