按用户组合周期

Question

我想把句号(start ->结束日期)，它们重叠成一个用户的间隔。

+---------+-------------+------------+------------+
| USER_ID | CONTRACT_ID | START_DATE |  END_DATE  |
+---------+-------------+------------+------------+
|       1 |          14 | 18.02.2021 | 18.04.2022 |
|       1 |          13 | 01.01.2019 | 01.01.2020 |
|       1 |          12 | 01.01.2018 | 01.01.2019 |
|       1 |          11 | 13.02.2017 | 13.02.2019 |
|       2 |          23 | 19.06.2021 | 18.04.2022 |
|       2 |          22 | 01.07.2019 | 01.07.2020 |
|       2 |          21 | 19.01.2019 | 19.01.2020 |
+---------+-------------+------------+------------+

作为结果，我想要这样的表:

+---------+------------+------------+
| USER_ID | START_DATE |  END_DATE  |
+---------+------------+------------+
|       1 | 18.02.2021 | 18.04.2022 |
|       1 | 13.02.2017 | 01.01.2020 |
|       2 | 19.06.2021 | 18.04.2022 |
|       2 | 19.01.2019 | 01.07.2020 |
+---------+------------+------------+

我尝试了不同的选项，但似乎都不起作用。

Answer 1

WITH DATAD (USER_ID,CONTRACT_ID,START_DATE,END_DATE)AS
(
SELECT 1,14,'2021-02-18','2022-04-18' UNION ALL
SELECT 1,13,'2019-01-01','2020-01-01' UNION ALL
SELECT 1,12,'2018-01-01','2019-01-01' UNION ALL
SELECT 1,11,'2017-02-13','2019-02-13' UNION ALL
SELECT 2,23,'2021-06-19','2022-04-18' UNION ALL
SELECT 2,22,'2019-07-01','2020-07-01' UNION ALL
SELECT 2,21,'2019-01-19','2020-01-19' 
),
CTE AS
(
SELECT USER_ID,START_DATE AS TS ,1 AS TYPE,
ROW_NUMBER()OVER(PARTITION BY USER_ID ORDER BY START_DATE) AS S,NULL AS E 
FROM DATAD
UNION ALL
SELECT USER_ID,END_DATE AS TS,-1 AS TYPE,
NULL,ROW_NUMBER()OVER(PARTITION BY USER_ID ORDER BY END_DATE)AS E
FROM DATAD
),
C2 AS
(
SELECT *,ROW_NUMBER()OVER(PARTITION BY USER_ID ORDER BY TS,TYPE DESC)AS SE 
FROM CTE 
),
C3 AS 
(
SELECT *,
FLOOR((ROW_NUMBER()OVER(PARTITION BY USER_ID ORDER BY TS)+1)/2)AS P
FROM C2
CROSS APPLY (VALUES(S-(SE-S)-1,(SE-E)-E)) AS A(CS,CE)
WHERE CS=0 OR CE=0
)
SELECT USER_ID,MIN(TS)AS START_TIME,MAX(TS)AS END_TIME
FROM C3
GROUP BY USER_ID,P
ORDER BY USER_ID;

这是一个"复制粘贴"摘自《T-SQL查询》一书。作者:Itzik Ben-Gan。关于"包装间隔"的段落

Answer 2

好的，所以我使用以下逻辑来获得答案-

1. 将范围的结束日期增加到重叠的范围
2. 完全消除位于另一个范围内的所有范围。

所以我使用了一个cte来绘制第一步，并使用WHERE NOT EXISTS作为第二步，就像这样

; with extendedset as 
(
select 
A.user_id,
A.contract_id, 
A.start_date, 
max(case when B.user_id is null then A.end_date else B.end_date end) end_date
from t A left join t B 
on 
A.user_id=B.user_id 
and A.contract_id<>B.contract_id
and A.end_date>=B.start_date
and A.end_date <=B.end_date
group by A.user_id,
A.contract_id, 
A.start_date
)
select 
user_id,start_date, end_date 
from extendedset t1
where not exists 
( select 1 from extendedset t2
where t1.user_id=t2.user_id and t1.contract_id<>t2.contract_id
and t1.start_date >= t2.start_date and t1.end_date <=t2.end_date)

参见working dbfiddle