我有一个分层产品价格列表,其价格已汇总。层级级别为群集,商店和货架。
我想按层次排序这些产品价格。每个级别都需要进行排序。以下是预期结果的示例:
现在,我可以在货架级别对价格进行排序,但不能在商店或集群级别进行排序:
SELECT cluster_id, store_id, shelf_id, price FROM (
SELECT * FROM (
SELECT p.*,
ROW_NUMBER () OVER (
ORDER BY
cluster_id ASC NULLS FIRST,
store_id ASC NULLS FIRST,
price DESC
) AS rn
FROM (SELECT * FROM products) p
)
)
ORDER BY rn;
我的实际结果是:
我正在使用Oracle Database 12c。
示例数据
CREATE TABLE products (
cluster_id VARCHAR2(30),
store_id VARCHAR2(30),
shelf_id VARCHAR2(40),
price NUMBER(*,2)
);
INSERT INTO products VALUES ('10','230',NULL,120);
INSERT INTO products VALUES (NULL,NULL,NULL,500);
INSERT INTO products VALUES ('10','230','967',40);
INSERT INTO products VALUES ('10',NULL,NULL,300);
INSERT INTO products VALUES ('50','430','863',50);
INSERT INTO products VALUES ('50','170',NULL,70);
INSERT INTO products VALUES ('10','500','783',100);
INSERT INTO products VALUES ('50','170','798',20);
INSERT INTO products VALUES ('50','480',NULL,80);
INSERT INTO products VALUES ('50','430',NULL,50);
INSERT INTO products VALUES ('10','500',NULL,180);
INSERT INTO products VALUES ('50','480','486',60);
INSERT INTO products VALUES ('10','230','296',80);
INSERT INTO products VALUES ('50',NULL,NULL,200);
INSERT INTO products VALUES ('10','500','344',80);
INSERT INTO products VALUES ('50','480','234',20);
INSERT INTO products VALUES ('50','170','368',50);
答案 0 :(得分:0)
我认为这会奏效。
<div class="field-panel">
<div>
<span>
<input #newData [(ngModel)]="data.id" type="text" >
<select #newCategory [(ngModel)]="i.index" class="input-bars">
<option *ngFor="let data of field" [value]="data.value">{{data.value}}</option>
</select>
</span>
<button class="btn btn-success (click)="addData(newData.value, newCategory.value)">ADD</button>
</div>
</div>
addData(value, dropValue){
this.Data[index] = value;
}
答案 1 :(得分:0)
这种排序方法提供了所需的输出:
select p.*
from products p
order by
sum(case when store_id is not null and shelf_id is not null then price end)
over (partition by cluster_id) desc,
sum(case when store_id is not null and shelf_id is not null then price end)
over (partition by cluster_id, store_id) desc,
case when shelf_id is null then 1 end, price desc