我使用此查询来汇总头寸的元素及其子元素,以检查价格是否正确。当我运行此查询时,我得到正确的结果,但是对于我的元素表超过2M +行和子元素9M +行,它的工作时间将近2分钟。
我认为由于子查询上的联接,MySQL无法正确使用索引,但我不知道如何解决此问题。
SELECT `pos`.`id`,
`pos`.`Name`,
`pos`.`Mark`,
`pos`.`Number`,
`elem`.`elemPrice` as `elemPrice`,
`elem`.`subelemPrice` as `subelemPrice`,
`pos`.`price` as `price`,
`elem`.`elemCalcPrice` as `elemCalcPrice`,
`elem`.`subelemCalcPrice` as `subelemCalcPrice`,
`pos`.`supplier_webcalcprice`
FROM `proposal_draft_positions` `pos`
LEFT JOIN (
SELECT SUM(quantity * price) as elemPrice,
SUM(quantity * supplier_webcalcprice) as elemCalcPrice,
SUM(quantity * subelem.subelemPrice) as subelemPrice,
SUM(quantity * subelem.subelemCalcPrice) as subelemCalcPrice,
position,
id,
quantity
FROM proposal_position_elements elm
LEFT JOIN (
SELECT SUM(price * quantity) as subelemPrice,
SUM(supplier_webcalcprice * quantity) as subelemCalcPrice,
element
FROM proposal_position_subelements
GROUP BY element) subelem ON subelem.element = elm.id
GROUP BY position) elem ON `elem`.`position` = `pos`.`id`
WHERE `draft_id` = 29407
EXPLAIN显示MySQL在连接时未正确使用索引:
表DDL:
create table proposal_draft_positions
(
id int(10) auto_increment
primary key,
Number int(10) null,
Mark varchar(200) null,
Name json null,
price decimal(10, 2) default 0.00 not null,
supplier_webcalcprice decimal(10, 2) default 0.00 null,
)
charset = utf8;
create table proposal_position_elements
(
id int(30) auto_increment primary key,
position int(10) not null,
quantity decimal(10, 2) default 1.00 not null,
price decimal(10, 2) default 0.00 null,
supplier_webcalcprice decimal(10, 2) default 0.00 null,
constraint fk_proposal_position_elements
foreign key (position) references proposal_draft_positions (id)
on delete cascade
)
charset = utf8;
create index `positions-elements-fk`
on proposal_position_elements (position);
create table proposal_position_subelements
(
id int(50) auto_increment
primary key,
element int(30) not null,
quantity decimal(10, 2) default 1.00 not null,
price decimal(10, 2) default 0.00 null,
supplier_webcalcprice decimal(10, 2) default 0.00 null,
constraint fk_proposal_position_subelements
foreign key (element) references proposal_position_elements (id)
on delete cascade
)
charset = utf8;
create index `element-subelement-fk`
on proposal_position_subelements (element);
正如EXPLAIN显示为整个表运行JOIN`查询一样,我决定添加过滤和小的重写查询:
SELECT `pos`.`id`,
`pos`.`Name`,
`pos`.`Mark`,
`pos`.`Number`,
`pos`.`supplier_webcalcprice`,
`pos`.`price` as `price`,
`elem`.`elemPrice` as `elemPrice`,
`elem`.`elemCalcPrice` as `elemCalcPrice`,
`elem`.`subelemPrice` as `subelemPrice`,
`elem`.`subelemCalcPrice` as `subelemCalcPrice`
FROM `proposal_draft_positions` `pos`
LEFT JOIN (
SELECT SUM(elem.price * elem.quantity) as elemPrice,
SUM(elem.supplier_webcalcprice * elem.quantity) as elemCalcPrice,
SUM(elem.quantity * subelem.subelemPrice) as subelemPrice,
SUM(elem.quantity * subelem.subelemCalcPrice) as subelemCalcPrice,
elem.position,
elem.id,
elem.quantity
FROM `proposal_draft_positions` pdp
LEFT JOIN proposal_position_elements elem on pdp.id = elem.position
LEFT JOIN (
SELECT SUM(subelem.price * subelem.quantity) as subelemPrice,
SUM(subelem.supplier_webcalcprice * subelem.quantity) as subelemCalcPrice,
element
FROM `proposal_draft_positions` pdp
LEFT JOIN proposal_position_elements elem on pdp.id = elem.position
LEFT JOIN proposal_position_subelements subelem on elem.id = subelem.element
WHERE pdp.draft_id = 29407
GROUP BY element) subelem ON subelem.element = elem.id
WHERE pdp.draft_id = 29407
GROUP BY position) elem ON `elem`.`position` = `pos`.`id`
WHERE `draft_id` = 29407;