Question

我似乎无法正确理解这一点。我需要基于总组中rev_amount的最高总和的客户排名，然后按bu_list中每个条目的收入总和排名 - 这可能吗？我在一个查询中想要这个的原因是因为我在下面制作了一个示例值列表，但我的实际表有20,000个客户和大约25个bu_id，所以我真的很简单。我在一个查询中一直在摆弄排名和支点，但似乎无法做到正确。谢谢你的帮助。

CREATE TABLE bu_list ( bu_id int, bu_name varchar(50) );
CREATE TABLE cust_list ( cust_id int, rev_amount int, bu_id_fk int );

INSERT INTO bu_list VALUES (1, 'East');
INSERT INTO bu_list VALUES (2, 'West');
INSERT INTO bu_list VALUES (3, 'North');
INSERT INTO bu_list VALUES (4, 'South');

INSERT INTO cust_list VALUES (1, 1000, 1);
INSERT INTO cust_list VALUES (2, 2000, 1);
INSERT INTO cust_list VALUES  (3, 2000, 2);
INSERT INTO cust_list VALUES  (4, 2000, 3);
INSERT INTO cust_list VALUES  (5, 4000, 3);
INSERT INTO cust_list VALUES  (5, 100, 3);
INSERT INTO cust_list VALUES  (6, 5000, 4);
INSERT INTO cust_list VALUES  (7, 6000, 4);

预期输出

cust_id |total_rank| east_rank| west_rank| north_rank| south_rank
7         1                                                1
6         2                                                2
5         3                                  1
4         4                                  2
3         4                         1
2         4              1
1         5              2

Answer 1

您可以使用条件聚合执行此操作：

SELECT
    cust_id,
    total_rank,
    east_rank   = MAX(CASE WHEN b.bu_id = 1 THEN t.bu_rank END),
    west_rank   = MAX(CASE WHEN b.bu_id = 2 THEN t.bu_rank END),
    north_rank  = MAX(CASE WHEN b.bu_id = 3 THEN t.bu_rank END),
    south_rank  = MAX(CASE WHEN b.bu_id = 4 THEN t.bu_rank END)
FROM (
    SELECT
        cust_id,
        bu_id_fk,
        total_rank  = DENSE_RANK() OVER(ORDER BY rev_amount DESC),
        bu_rank     = DENSE_RANK() OVER(PARTITION BY bu_id_fk ORDER BY rev_amount DESC)
    FROM cust_list
) t
INNER JOIN bu_list b
    ON b.bu_id = t.bu_id_fk
GROUP BY t.cust_id, t.total_rank
ORDER BY t.total_rank

具有rank和pivot的sql - 需要在每个分区中排名并在一个查询中排名

1 个答案: