Question

想象一下，我有下表（动物）：

**Color**       **Species**    **Weight**
White              Dog            20
White              Dog            8
White              Dog            33
Black              Dog            55
Brown              Dog            80
White              Cat            10
Black              Cat            14
White              Cat            9

我希望按物种分组，过滤每个物种内的独特颜色，并为每个过滤组找到两个最轻的动物。

结果表应如下所示：

**Color**       **Species**   **Weight**  
White              Dog            8         
Black              Dog            55
White              Cat            9
Black              Cat            14

我使用以下查询（我知道这是不正确的）：

SELECT color, species, weight
FROM (
    SELECT species, color, weight, rank()
           over (PARTITION BY species ORDER BY weight ASC) as rank
    FROM animals) ranked_animals
WHERE ranked_animals.rank <= 2;

我不知道如何过滤上述代码中的唯一颜色。有什么建议吗？

谢谢！

Answer 1

样本表

CREATE TABLE #TEMP([COLOR] VARCHAR(20),Species VARCHAR(20), [Weight] INT)

INSERT INTO #TEMP
SELECT 'White' ,             'Dog',            20
UNION ALL
SELECT 'White',              'Dog',            8
UNION ALL
SELECT 'White',              'Dog',            33
UNION ALL
SELECT 'Black' ,             'Dog',            55
UNION ALL
SELECT 'Brown' ,             'Dog',            80
UNION ALL
SELECT 'White',              'Cat',            10
UNION ALL
SELECT 'Black',              'Cat',            14
UNION ALL
SELECT 'White' ,            'Cat' ,          9

<强> QUERY

;WITH CTE AS
(
    -- First partition with [COLOR],Species and generate ROW_NUMBER
    SELECT DISTINCT  [COLOR],Species,[Weight],
    ROW_NUMBER() OVER (PARTITION BY [COLOR],Species ORDER BY [Weight] ) RNO 
    FROM #TEMP  
)
,CTE2 AS
(
    -- Next  partition with Species only and generate ROW_NUMBER
    SELECT *,ROW_NUMBER() OVER (PARTITION BY Species ORDER BY [Weight] ) RNO2
    FROM CTE
    WHERE RNO = 1 
)
-- Now take new ROW_NUMBER() ie, RNO2 <= 2
SELECT [COLOR],Species,[Weight]
FROM CTE2 
WHERE RNO2< = 2
ORDER BY Species DESC,[COLOR] DESC

Click here查看结果

使用PARTITION BY（HIVE）时如何过滤掉组中的重复元素

1 个答案: