当车辆在特定日期加油时，我试图获得之前的燃油消耗。
我的表格结构如下：
id vehicle center dates odometer quantity
----------------------------------------------------------------------------
11 UAY329R 257 2017-04-01 153329 15
12 S022 254 2017-03-22 139828 15
13 UAP614Z 254 2017-04-01 211410 15
14 S022 254 2017-03-25 139928 15
15 UAY848Q 254 2017-04-01 239813 15
16 C052 257 2017-04-11 15016 12.06
17 UAP258A 254 2017-03-29 29495 20
18 S022 254 2017-04-12 140078 35.66
19 UAP258A 254 2017-04-01 29575 20
20 UDL146E 254 2017-04-01 223701 20
21 UAP258A 254 2017-04-03 29675 20
22 UDL146E 254 2017-04-04 223851 5
prev_id PrevOdo CurrentId CurrentOdo Distance Vehicle Dates
----------------------------------------------------------------------------------------
8 139708 12 139828 120 S022 2017-03-22
12 139828 14 139928 100 S022 2017-03-25
12 139928 18 140078 150 S022 2017-04-12
10 29415 17 29495 80 UAP258A 2017-03-29
17 29495 19 29575 80 UAP258A 2017-04-01
19 29575 21 29675 100 UAP258A 2017-04-03
7 223601 20 223701 100 UDL146E 2017-04-01
20 223701 22 223851 150 UDL146E 2017-04-04
我的查询部分有效，但如果以前的交易中有多条记录，则第二条记录显示错误的PrevOdo。
我的查询如下所示，请帮助我，并提前感谢您。
SELECT prev_id, p.odometer AS 'PrevOdo', t.id AS 'CurrentId', t.odometer AS 'CurrentOdo', (t.odometer - p.odometer) AS 'Distance', t.vehicle AS 'Vehicle', t.dates AS 'Dates'
FROM ( SELECT t.id, t.odometer, t.vehicle, t.dates, t.quantity, (
SELECT id
FROM ISSUANCE
WHERE dates < t.dates AND vehicle = t.vehicle
ORDER BY id DESC
LIMIT 1
) prev_id
FROM ISSUANCE t
WHERE t.dates BETWEEN '2017-04-01' AND '2017-04-15')
t LEFT JOIN ISSUANCE p ON t.prev_id = p.id
ORDER BY t.vehicle
查询按日期过滤。
这是我用过的表格结构因为某些记录混淆了所以我不能跟随id，使用日期。提前谢谢。
drop table if exists issuance;
create table issuance (
id int,
vehicle varchar(20),
center int,
dates date,
odometer int,
quantity int
);
insert into issuance values
(02, 'S022', 254, '2017-04-14', 140178, 5),
(04, 'S022', 254, '2017-04-18', 140378, 5),
(11, 'UAY329R', 257, '2017-04-01', 153329, 15),
(12, 'S022', 254, '2017-03-22', 139828, 15),
(13, 'UAP614Z', 254, '2017-04-01', 211410, 15),
(14, 'S022', 254, '2017-03-25', 139928, 15),
(15, 'UAY848Q', 254, '2017-04-01', 239813, 15),
(16, 'C052', 257, '2017-04-11', 15016 , 12.06),
(17, 'UAP258A', 254, '2017-03-29', 29495 , 20),
(18, 'S022', 254, '2017-04-12', 140078, 35.66),
(19, 'UAP258A', 254, '2017-04-01', 29575 , 20),
(20, 'UDL146E', 254, '2017-04-01', 223701, 20),
(21, 'UAP258A', 254, '2017-04-03', 29675 , 20),
(22, 'UDL146E', 254, '2017-04-04', 223851, 5),
(23, 'S022', 254, '2017-03-26', 139948, 5),
(25, 'S022', 254, '2017-04-16', 140278, 5),
(27, 'S022', 254, '2017-04-19', 140478, 5),
(98, 'S022', 254, '2017-04-22', 140578, 5);
select prev.id as PrevId,
prev.odometer as PrevOdo,
cur.id as CurId,
cur.odometer as CurOdo,
cur.odometer - prev.odometer as distance,
cur.vehicle as Vehicle,
cur.dates as Dates
from issuance cur
join (
select t1.id, max(t2.id) as prev_id
from issuance t1
join issuance t2
on t1.vehicle = t2.vehicle and
t1.dates > t2.dates
group by t1.id
) mid
on cur.id = mid.id
join issuance prev
on prev.id = mid.prev_id
where cur.dates between '2017-03-01' and '2017-05-15'
ORDER BY Vehicle ASC, Dates ASC;
答案 0 :(得分：0)
我会分两步建立一个查询：首先我会得到id / prev_id夫妻
select t1.id, max(t2.id) as prev_id
from issuance t1
join issuance t2
on t1.vehicle = t2.vehicle and
t1.id > t2.id
group by t1.id
现在我将它用作原始表的自连接中的中间连接：
select prev.id as PrevId,
prev.odometer as PrevOdo,
cur.id as CurId,
cur.odometer as CurOdo,
cur.odometer - prev.odometer as distance,
cur.vehicle as Vehicle,
cur.dates as Dates
from issuance cur
join (
select t1.id, max(t2.id) as prev_id
from issuance t1
join issuance t2
on t1.vehicle = t2.vehicle and
t1.id > t2.id
group by t1.id
) mid
on cur.id = mid.id
join issuance prev
on prev.id = mid.prev_id
where cur.dates between '2017-04-01' and '2017-04-15';
修改
由于您不能依赖ID排序与日期排序一致，因此您必须更改加入条件;首先，您需要内部查询的最新日期而不是最大ID：
select t1.id, max(t2.dates) as prev_dates
from issuance t1
join issuance t2
on t1.vehicle = t2.vehicle and
t1.dates > t2.dates
group by t1.id
然后这个表必须加入不同的
select prev.id as PrevId,
prev.odometer as PrevOdo,
cur.id as CurId,
cur.odometer as CurOdo,
cur.odometer - prev.odometer as distance,
cur.vehicle as Vehicle,
cur.dates as Dates
from issuance cur
join (
select t1.id, max(t2.dates) as prev_dates
from issuance t1
join issuance t2
on t1.vehicle = t2.vehicle and
t1.dates > t2.dates
group by t1.id
) mid
on cur.id = mid.id
join issuance prev
on prev.dates = mid.prev_dates and
prev.vehicle = cur.vehicle /* the mid calculated field is not sufficient anymore */
where cur.dates between '2017-03-01' and '2017-05-15'
order by cur.vehicle asc, cur.dates asc
工作rextester here