JOIN内用FETCH FIRST ROW ONLY获NULL值及SQL性能与分组优化问询
SQL关联查询性能与分组优化问题
之前我问过在JOIN语句中引用外部变量的问题,虽得到解决方案,但执行耗时超一分钟,性能极差。
示例表结构
TIMESIN表
| ID | TIME_IN |
|---|---|
| 1 | 30-JUN-2023 12:00:00 |
| 1 | 29-JUN-2023 08:00:00 |
| 1 | 28-JUN-2023 03:00:00 |
| 1 | 27-JUN-2023 19:00:00 |
| 2 | 30-JUN-2023 07:00:00 |
| 2 | 29-JUN-2023 12:00:00 |
MOVEMENTTAB表
| ID | TIME_STAMP | COMMAND |
|---|---|---|
| 1 | 30-JUN-2023 12:10:00 | 2 |
| 1 | 29-JUN-2023 08:10:00 | 2 |
| 1 | 28-JUN-2023 03:10:00 | 2 |
| 1 | 27-JUN-2023 19:10:00 | 2 |
| 1 | 26-JUN-2023 08:10:00 | 2 |
初始SQL与结果
执行以下SQL可得到所有大于TIME_IN的TIME_OUT,结果符合预期:
SELECT A.ID, A."TIME_IN", MOVE.TIME_OUT FROM TIMESIN A LEFT JOIN ( SELECT ID, TIME_STAMP AS "TIME_OUT" FROM MOVEMENTTAB M WHERE COMMAND = 2 ) MOVE ON MOVE.ID = A.ID AND MOVE.TIME_OUT > A.TIME_IN
查询结果:
| ID | TIME_IN | TIME_OUT |
|---|---|---|
| 1 | 30-JUN-2023 12:00:00 | 30-JUN-2023 12:10:00 |
| 1 | 29-JUN-2023 08:00:00 | 29-JUN-2023 08:10:00 |
| 1 | 29-JUN-2023 08:00:00 | 30-JUN-2023 12:10:00 |
| 1 | 28-JUN-2023 03:00:00 | 28-JUN-2023 03:10:00 |
| 1 | 28-JUN-2023 03:00:00 | 29-JUN-2023 08:10:00 |
| 1 | 28-JUN-2023 03:00:00 | 30-JUN-2023 12:10:00 |
| 1 | 27-JUN-2023 19:00:00 | 27-JUN-2023 19:10:00 |
| 1 | 27-JUN-2023 19:00:00 | 28-JUN-2023 03:10:00 |
| 1 | 27-JUN-2023 19:00:00 | 29-JUN-2023 08:10:00 |
| 1 | 27-JUN-2023 19:00:00 | 30-JUN-2023 12:10:00 |
遇到的问题
- 在
LEFT JOIN的子查询中添加FETCH FIRST ROW ONLY和ORDER BY TIME_OUT后,TIME_OUT字段出现NULL值。 - 测试过其他方案(包括
CROSS JOIN LATERAL),性能依旧很差,耗时同样超一分钟。
分组优化的疑问
考虑先执行上述初始SQL,再通过GROUP BY TIME_IN和MIN(MOVE.TIME_OUT)来获取每个TIME_IN对应的最早TIME_OUT。但如果TIMESIN表有更多需要保留的列,是否有更简洁的写法?比如MIN(A.*)这种方式?还是必须为每个列写MIN(A.x),或者在GROUP BY中列出所有A.x, A.y这类列?另外,这种分组写法是否存在性能问题?
内容的提问来源于stack exchange,提问作者FuzzUK
相关产品推荐
相关产品推荐

