使用Starrocks制作拉链表

发布于:2025-05-23 ⋅ 阅读:(16) ⋅ 点赞:(0)

5月1日向ods_order_info插入3条数据:

CREATE TABLE ods_order_info(
  `dt` string,
  `id` string COMMENT '订单编号',
  `total_amount` decimal(10,2) COMMENT '订单金额'
)
PRIMARY KEY(`dt`, `id`)
PARTITION BY (`dt`)
DISTRIBUTED BY HASH(`id`)
PROPERTIES (
"replication_num" = "1"
);

insert into ods_order_info
select 
    '2025-05-01'          as dt
    ,9527                 as id
    ,2000                 as total_amount
;

insert into ods_order_info
select 
    '2025-05-01'          as dt
    ,9528                 as id
    ,3000                 as total_amount
;

insert into ods_order_info
select 
    '2025-05-01'          as dt
    ,9529                 as id
    ,4000                 as total_amount
;

查询结果:

MySQL [tmp]> select * from ods_order_info;
+------------+------+--------------+
| dt         | id   | total_amount |
+------------+------+--------------+
| 2025-05-01 | 9529 |      4000.00 |
| 2025-05-01 | 9528 |      3000.00 |
| 2025-05-01 | 9527 |      2000.00 |
+------------+------+--------------+
3 rows in set (0.01 sec)

制作拉链表dwd_order_info_his:

drop table if exists dwd_order_info_his;
create table dwd_order_info_his( 
  `end_date`  string COMMENT '有效结束日期',
  `id` string COMMENT '订单编号',
  `total_amount` decimal(10,2) COMMENT '订单金额', 
  `start_date`  string COMMENT '有效开始日期',
  `record_status` string COMMENT '是否有效'
)
PRIMARY KEY(`end_date`, `id`)
PARTITION BY (`end_date`)
DISTRIBUTED BY HASH(`id`)
PROPERTIES (
"replication_num" = "1"
);

初始化拉链表:

insert overwrite dwd_order_info_his
select
    '9999-99-99',
    id,
    total_amount,
    '2025-05-01',
    'active'
from ods_order_info di
where di.dt='2025-05-01';

查询结果:

MySQL [tmp]> select * from dwd_order_info_his;
+------------+------+--------------+------------+---------------+
| end_date   | id   | total_amount | start_date | record_status |
+------------+------+--------------+------------+---------------+
| 9999-99-99 | 9527 |      2000.00 | 2025-05-01 | active        |
| 9999-99-99 | 9529 |      4000.00 | 2025-05-01 | active        |
| 9999-99-99 | 9528 |      3000.00 | 2025-05-01 | active        |
+------------+------+--------------+------------+---------------+
3 rows in set (0.01 sec)

创建临时表用于导数据:

drop table if exists dwd_order_info_his_tmp;
create table dwd_order_info_his_tmp( 
  `end_date`  string COMMENT '有效结束日期',
  `id` string COMMENT '订单编号',
  `total_amount` decimal(10,2) COMMENT '订单金额', 
  `start_date`  string COMMENT '有效开始日期'
)
PRIMARY KEY(`end_date`, `id`)
PARTITION BY (`end_date`)
DISTRIBUTED BY HASH(`id`)
PROPERTIES (
"replication_num" = "1"
);

5月2日ODS表发生改变:

insert into ods_order_info
select 
    '2025-05-02'          as dt
    ,9527                 as id
    ,2222                 as total_amount
;

insert into ods_order_info
select 
    '2025-05-02'          as dt
    ,9540                 as id
    ,7000                 as total_amount
;

导入数据:

insert overwrite dwd_order_info_his_tmp
select * from 
(
select 
    '9999-99-99' end_date,
    id,
    total_amount,
    '2025-05-02' start_date
from ods_order_info where dt='2025-05-02'
union all 
select 
    if(oi.id is null, oh.end_date, date_add(oi.dt, -1)) end_date,
    oh.id,
    oh.total_amount,
    oh.start_date
from dwd_order_info_his oh left join 
     (
      select
      *
      from ods_order_info
      where dt='2025-05-02'
) oi
     on oh.id=oi.id and oh.end_date='9999-99-99'  
)his 
order by his.id, start_date;


insert overwrite dwd_order_info_his 
select 
end_date,
id,
total_amount,
start_date,
case when end_date = '9999-99-99' then 'active' else 'expire' end as record_status
from dwd_order_info_his_tmp;

查询结果:

MySQL [tmp]> select * from dwd_order_info_his where start_date = '2025-05-01' and record_status = 'active';
+------------+------+--------------+------------+---------------+
| end_date   | id   | total_amount | start_date | record_status |
+------------+------+--------------+------------+---------------+
| 9999-99-99 | 9529 |      4000.00 | 2025-05-01 | active        |
| 9999-99-99 | 9528 |      3000.00 | 2025-05-01 | active        |
+------------+------+--------------+------------+---------------+
2 rows in set (0.01 sec)

MySQL [tmp]> select * from dwd_order_info_his where start_date = '2025-05-02' and record_status = 'active';
+------------+------+--------------+------------+---------------+
| end_date   | id   | total_amount | start_date | record_status |
+------------+------+--------------+------------+---------------+
| 9999-99-99 | 9527 |      2222.00 | 2025-05-02 | active        |
| 9999-99-99 | 9540 |      7000.00 | 2025-05-02 | active        |
+------------+------+--------------+------------+---------------+
2 rows in set (0.01 sec)

5月3日ODS层数据再次改变:

insert into ods_order_info
select 
    '2025-05-03'          as dt
    ,9528                 as id
    ,3333                 as total_amount
;

insert into ods_order_info
select 
    '2025-05-03'          as dt
    ,9541                 as id
    ,8000                 as total_amount
;

导入数据:

insert overwrite dwd_order_info_his_tmp
select * from 
(
select 
    '9999-99-99' end_date,
    id,
    total_amount,
    '2025-05-03' start_date
from ods_order_info where dt='2025-05-03'
union all 
select 
    if(oi.id is null, oh.end_date, date_add(oi.dt, -1)) end_date,
    oh.id,
    oh.total_amount,
    oh.start_date
from dwd_order_info_his oh left join 
     (
      select
      *
      from ods_order_info
      where dt='2025-05-03'
) oi
     on oh.id=oi.id and oh.end_date='9999-99-99'  
)his 
order by his.id, start_date;

查询数据:

MySQL [tmp]> select * from dwd_order_info_his;
+---------------------+------+--------------+------------+---------------+
| end_date            | id   | total_amount | start_date | record_status |
+---------------------+------+--------------+------------+---------------+
| 9999-99-99          | 9529 |      4000.00 | 2025-05-01 | active        |
| 9999-99-99          | 9541 |      8000.00 | 2025-05-03 | active        |
| 2025-05-01 00:00:00 | 9527 |      2000.00 | 2025-05-01 | expire        |
| 9999-99-99          | 9528 |      3333.00 | 2025-05-03 | active        |
| 9999-99-99          | 9540 |      7000.00 | 2025-05-02 | active        |
| 9999-99-99          | 9527 |      2222.00 | 2025-05-02 | active        |
| 2025-05-02 00:00:00 | 9528 |      3000.00 | 2025-05-01 | expire        |
+---------------------+------+--------------+------------+---------------+
7 rows in set (0.01 sec)

MySQL [tmp]> select * from dwd_order_info_his where start_date = '2025-05-01' and record_status = 'active';
+------------+------+--------------+------------+---------------+
| end_date   | id   | total_amount | start_date | record_status |
+------------+------+--------------+------------+---------------+
| 9999-99-99 | 9529 |      4000.00 | 2025-05-01 | active        |
+------------+------+--------------+------------+---------------+
1 row in set (0.03 sec)

MySQL [tmp]> select * from dwd_order_info_his where start_date = '2025-05-02' and record_status = 'active';
+------------+------+--------------+------------+---------------+
| end_date   | id   | total_amount | start_date | record_status |
+------------+------+--------------+------------+---------------+
| 9999-99-99 | 9540 |      7000.00 | 2025-05-02 | active        |
| 9999-99-99 | 9527 |      2222.00 | 2025-05-02 | active        |
+------------+------+--------------+------------+---------------+
2 rows in set (0.01 sec)

MySQL [tmp]> select * from dwd_order_info_his where start_date = '2025-05-03' and record_status = 'active';
+------------+------+--------------+------------+---------------+
| end_date   | id   | total_amount | start_date | record_status |
+------------+------+--------------+------------+---------------+
| 9999-99-99 | 9541 |      8000.00 | 2025-05-03 | active        |
| 9999-99-99 | 9528 |      3333.00 | 2025-05-03 | active        |
+------------+------+--------------+------------+---------------+
2 rows in set (0.01 sec)


网站公告

今日签到

点亮在社区的每一天
去签到