非唯一列的选择和计数属性

Question

I have a table of persons and activities - neither column is unique. I need to rank every user by the count of distinct activities, eg:

_________________ 
|PERSON|ACTIVITY|
-----------------
|Lars  | Sleep  |
|James | Eat    |
|Lars  | Sleep  | 
|Lars  | Sleep  | 
|Kirk  | Shred  |
|James | Shred  |
-----------------

Lars appears thrice, but performs the same activity repeatedly. Kirk appears once, so he is identical to Lars in number of activities. James performs two distinct activities, so he should be ranked the highest.

The expected output:

James - 2
Kirk - 1 
Lars - 1

(ordering of identical counts is irrelevant)

The solution I have come up with involves applying DISTINCT to the person column and iterating over the names, selecting the activities for each and applying DISTINCT followed by COUNT. It feels like there must be a better way.

Answer 1

I think you just want count(distinct) :

select person, count(distinct activity) as num_activities
from t
group by person
order by num_activities desc;

Answer 2

You could use GROUP BY function. SELECT PERSON, COUNT(DISTINCT ACTIVITY) AS count FROM YOUR TABLE GROUP BY PERSON ORDER BY count;

Answer 3

SELECT Person, COUNT(DISTINCT Activity) AS ActivityCount
FROM MyTable
GROUP BY Person
ORDER BY 2 DESC