sql - 如何每周获取前 10 名数据（本周、前一周、上个月、2 个月前、3 个月前）

Question

我想检索 Top 10 产品 - 每周

Rank     This week product   Previous week rank   Last month rank   2 month ago rank     3 month ago rank
 1             Watch                  2                    3                    1               4
 2             Radio                  3                    2                    4               5
 3             Pen                    4                    5                    6               7
 4
 5
 6
 7
 8
 9
 10

请帮助我了解如何从 SQL 查询中获取此类数据。谢谢

Answer 1

假设你有一张桌子：

CREATE TABLE your_table (
  product          VARCHAR2(20),
  some_date_column DATE
)

RANK然后您可以在分析函数中使用条件聚合：

SELECT product,
       RANK() OVER (
         ORDER BY COUNT(
           CASE
           WHEN some_date_column >= TRUNC(SYSDATE, 'IW')
           AND  some_date_column <  TRUNC(SYSDATE, 'IW') + INTERVAL '7' DAY
           THEN 1
           END
         ) DESC
       ) AS rank_this_week,
       RANK() OVER (
         ORDER BY COUNT(
           CASE
           WHEN some_date_column >= TRUNC(SYSDATE, 'IW') - INTERVAL '7' DAY
           AND  some_date_column <  TRUNC(SYSDATE, 'IW')
           THEN 1
           END
         ) DESC
       ) AS rank_last_week,
       RANK() OVER (
         ORDER BY COUNT(
           CASE
           WHEN some_date_column >= ADD_MONTHS(TRUNC(SYSDATE, 'IW'), -1)
           AND  some_date_column <  TRUNC(SYSDATE, 'MM')
           THEN 1
           END
         ) DESC
       ) AS rank_last_month,
       RANK() OVER (
         ORDER BY COUNT(
           CASE
           WHEN some_date_column >= ADD_MONTHS(TRUNC(SYSDATE, 'IW'), -2)
           AND  some_date_column <  ADD_MONTHS(TRUNC(SYSDATE, 'IW'), -1)
           THEN 1
           END
         ) DESC
       ) AS rank_two_month,
       RANK() OVER (
         ORDER BY COUNT(
           CASE
           WHEN some_date_column >= ADD_MONTHS(TRUNC(SYSDATE, 'IW'), -3)
           AND  some_date_column <  ADD_MONTHS(TRUNC(SYSDATE, 'IW'), -2)
           THEN 1
           END
         ) DESC
       ) AS rank_three_month
FROM   your_table
WHERE  some_date_column < TRUNC(SYSDATE, 'IW') + INTERVAL '7' DAY
AND    some_date_column >= ADD_MONTHS(TRUNC(SYSDATE, 'IW'), -3)
GROUP BY product
ORDER BY rank_this_week
FETCH FIRST 10 ROWS WITH TIES;

其中，对于样本数据：

INSERT INTO your_table (product, some_date_column)
SELECT 'A', TRUNC(SYSDATE, 'IW') FROM DUAL CONNECT BY LEVEL <=  1 UNION ALL
SELECT 'B', TRUNC(SYSDATE, 'IW') FROM DUAL CONNECT BY LEVEL <=  2 UNION ALL
SELECT 'C', TRUNC(SYSDATE, 'IW') FROM DUAL CONNECT BY LEVEL <=  3 UNION ALL
SELECT 'D', TRUNC(SYSDATE, 'IW') FROM DUAL CONNECT BY LEVEL <=  4 UNION ALL
SELECT 'E', TRUNC(SYSDATE, 'IW') FROM DUAL CONNECT BY LEVEL <=  5 UNION ALL
SELECT 'F', TRUNC(SYSDATE, 'IW') FROM DUAL CONNECT BY LEVEL <=  6 UNION ALL
SELECT 'G', TRUNC(SYSDATE, 'IW') FROM DUAL CONNECT BY LEVEL <=  7 UNION ALL
SELECT 'H', TRUNC(SYSDATE, 'IW') FROM DUAL CONNECT BY LEVEL <=  8 UNION ALL
SELECT 'I', TRUNC(SYSDATE, 'IW') FROM DUAL CONNECT BY LEVEL <=  9 UNION ALL
SELECT 'J', TRUNC(SYSDATE, 'IW') FROM DUAL CONNECT BY LEVEL <= 10 UNION ALL
SELECT 'K', TRUNC(SYSDATE, 'IW') FROM DUAL CONNECT BY LEVEL <= 11 UNION ALL
SELECT 'A', TRUNC(SYSDATE, 'IW') - INTERVAL '7' DAY FROM DUAL CONNECT BY LEVEL <=  3 UNION ALL
SELECT 'B', TRUNC(SYSDATE, 'IW') - INTERVAL '7' DAY FROM DUAL CONNECT BY LEVEL <=  3 UNION ALL
SELECT 'C', TRUNC(SYSDATE, 'IW') - INTERVAL '7' DAY FROM DUAL CONNECT BY LEVEL <=  1 UNION ALL
SELECT 'D', TRUNC(SYSDATE, 'IW') - INTERVAL '7' DAY FROM DUAL CONNECT BY LEVEL <=  2 UNION ALL
SELECT 'E', TRUNC(SYSDATE, 'IW') - INTERVAL '7' DAY FROM DUAL CONNECT BY LEVEL <=  5 UNION ALL
SELECT 'F', TRUNC(SYSDATE, 'IW') - INTERVAL '7' DAY FROM DUAL CONNECT BY LEVEL <=  7 UNION ALL
SELECT 'G', TRUNC(SYSDATE, 'IW') - INTERVAL '7' DAY FROM DUAL CONNECT BY LEVEL <=  6 UNION ALL
SELECT 'H', TRUNC(SYSDATE, 'IW') - INTERVAL '7' DAY FROM DUAL CONNECT BY LEVEL <=  4 UNION ALL
SELECT 'I', TRUNC(SYSDATE, 'IW') - INTERVAL '7' DAY FROM DUAL CONNECT BY LEVEL <=  5 UNION ALL
SELECT 'J', TRUNC(SYSDATE, 'IW') - INTERVAL '7' DAY FROM DUAL CONNECT BY LEVEL <=  3 UNION ALL
SELECT 'K', TRUNC(SYSDATE, 'IW') - INTERVAL '7' DAY FROM DUAL CONNECT BY LEVEL <=  2
-- Etc.

（注意：这只有过去 2 周的数据，所以所有其他排名都是平局。）

输出：

产品	RANK_THIS_WEEK	RANK_LAST_WEEK	RANK_LAST_MONTH	RANK_TWO_MONTH	RANK_THREE_MONTH
ķ	1	9	1	1	1
Ĵ	2	6	1	1	1
我	3	3	1	1	1
H	4	5	1	1	1
G	5	2	1	1	1
F	6	1	1	1	1
乙	7	3	1	1	1
D	8	9	1	1	1
C	9	11	1	1	1
乙	10	6	1	1	1

db<>在这里摆弄