Flagging repeated values

Question

Hi guys,
I'm wondering whether there is a better way to do this.
I have some IDs, and some scores, and then I rank them (based on the score) using the rank function. Two scores can have the same rank.
I want to flag whether or not a user is tied with another user (ie has the same rank) with either 'Y' or 'N'.
Here is an example table:
ID SCORE
1..... 22
2..... 43
3..... 43
4..... 54
5..... 65
SELECT r1.id,
       r1.score,
       r1.rank,
       CASE WHEN r2.rank_count > 1
            THEN 'Y'
            ELSE 'N'
       END tied
FROM
(SELECT
    id,
    score,
    RANK () OVER (ORDER BY score) rank
FROM test) r1,
(SELECT rank,
        count(rank) as rank_count
FROM
(SELECT
    id,
    score,
    RANK () OVER (ORDER BY score) rank
FROM test)
GROUP BY rank) r2
WHERE r1.rank = r2.rank
OUTPUT:
ID SCORE RANK TIED
1... 22...... 1..... N
3... 43...... 2..... Y
2... 43...... 2..... Y
4... 54...... 4..... N
5... 65...... 5..... N
What's the best way to go about doing this without the use of pseudo tables? Is there another analytic function I could use?

Answer 1

from your previous example,
SQL>
SQL>  With t As
  2      (  Select 1 item_id,52 rating From dual Union All
  3         Select 2        ,44        From dual Union All
  4         Select 3        ,32        From dual Union All
  5         Select 4        ,32        From dual Union All
  6         Select 5        ,20        From dual
  7      ) Select item_id,
  8               rating,
  9               rank,
10               Case When lead_rank = rank Or lag_rank = rank Then 'Y'
11               Else 'N'
12               End tied
13        From  (Select item_id,
14                      rating,
15                      rank,
16                      lead(rank) over (Order By rating Desc) lead_rank,
17                      lag(rank) over (Order By rating Desc)  lag_rank
18               From (Select item_id,
19                            rating,
20                            dense_rank() over (Order By rating Desc) rank
21                     From t
22                     )
23               )
24        Order By 1;
   ITEM_ID     RATING       RANK TIED
         1         52          1 N
         2         44          2 N
         3         32          3 Y
         4         32          3 Y
         5         20          4 N
SQL>