3

If I have a DB table called FAVORITE_FLAVOR where each row has a user's favorite flavor of ice cream.

User ID | Flavor      | State
1       | Chocolate   | CA
2       | Vanilla     | ND
3       |   Chocolate | CA
4       | Rocky Road  | CA
5       | vanilla     | CA
6       | Vanilla     | CA
7       | Vanilla     | CA

Now, if I want to query the 2 most popular flavors in each state (normalizing capitalization and whitespace), I could query:

SELECT state, INITCAP(TRIM(flavor)), count(INITCAP(TRIM(flavor))) AS total
FROM favorite_flavor GROUP BY state, INITCAP(TRIM(flavor))
ORDER BY state ASC, total DESC;

Which returns:

CA | Vanilla    | 3
CA | Chocolate  | 2
CA | Rocky Road | 1
ND | Vanilla    | 1

Now, I only wanted to know the top 2 flavors per state. How do I limit the query so that Rocky Road is no longer listed.

4

1 回答 1

5
SELECT
   State,
   flv,
   total
FROM (SELECT
         ROW_NUMBER() OVER ( PARTITION BY state ORDER BY count(INITCAP(TRIM(flavor))) DESC ) RowNumber,
         State,
         INITCAP(TRIM(flavor)) flv,
         count(INITCAP(TRIM(flavor))) total
      FROM favorite_flavor
      GROUP BY state, INITCAP(TRIM(flavor))
      ) dt
WHERE RowNumber <= 2
ORDER BY state ASC, total DESC
于 2013-06-27T23:57:55.853 回答