我有两个表Cars
和Defect
。我通过将car_id
作为外键传递到缺陷表来保持关系。
我在汽车表中有3000条记录,在缺陷表中有16000条记录(具有open_defect
和close_defect
)。我试图找出所有具有未解决缺陷的Cars(3000)(如果没有针对car的未解决缺陷,则应返回0)。
我正在尝试一些查询:
SELECT cars.cars_id_primary ,IF(COUNT(defect.defect_id_primary)>0,1,0) AS `def_count`
FROM cars
LEFT JOIN defect ON cars.cars_id_primary = defect.cars_id AND defect.defect_status_id =1
WHERE cars.stage_id !=5
GROUP BY cars.cars_id_primary
ORDER BY cars.updated_on
此查询给我结果,但是执行时间太长。 需要优化此查询。我坚持要优化。
欢迎任何帮助,谢谢。
答案 0 :(得分:2)
无需在GROUP BY
内加入汽车表。像这样重写查询:
SELECT cars.cars_id_primary, COALESCE(agg.open_defect_count, 0) AS open_defect_count
FROM cars
LEFT JOIN (
SELECT cars_id, COUNT(*) AS open_defect_count
FROM defect
WHERE defect_status_id = 1
GROUP BY cars_id
) AS agg ON cars.cars_id_primary = agg.cars_id
WHERE cars.stage_id != 5
ORDER BY cars.updated_on
您还将需要创建索引。我建议从ix_defect(defect_status_id, cars_id)
开始。
答案 1 :(得分:1)
您可以通过使用索引来提高性能。是否可以像下面的sql一样在cars_id_primary
表的cars
上创建索引并在cars_id
的{{1}}上创建索引。然后,您可以尝试查询。
defect
答案 2 :(得分:0)
我无法理解您的问题,我对查询进行了一些更改,
SELECT cars.cars_id_primary ,COUNT(CASE WHEN (defect.defect_id_primary)>0 THEN 1 ELSE 0
END) AS `def_count`
FROM cars
LEFT JOIN defect ON cars.cars_id_primary = defect.cars_id
WHERE cars.stage_id !=5 AND defect.defect_status_id =1
GROUP BY cars.cars_id_primary
ORDER BY cars.updated_on
尝试