sql模糊关联匹配

需求目标：

建立临时表

drop table grafana_bi.zbj_gift_2024;

USE grafana_bi;
CREATE  TABLE zbj_gift_2024 (
    id INT AUTO_INCREMENT PRIMARY KEY,
    userName VARCHAR(255),
    giftName VARCHAR(255),
    giftNum INT,
    points INT,
    teacher VARCHAR(255),
    sendDate DATETIME,
    terraceId INT,
    round_beginTime DATETIME,
    round_endTime DATETIME,
    round_teacher VARCHAR(255),
    ter INT
);

将查询数据插入临时表

-- V2.0 doubao AI ---------------------------------------------------------------------------

INSERT INTO grafana_bi.zbj_gift_2024 (id, userName, giftName, giftNum, points, teacher, sendDate, terraceId, round_beginTime, round_endTime, round_teacher, ter)
SELECT NULL, t1.userName, t1.giftName, t1.giftNum, t1.points, t1.teacher, t1.sendDate, t1.terraceId, 
       STR_TO_DATE(t2.beginTime, '%Y-%m-%d %H:%i:%s') beginTime, 
       STR_TO_DATE(t2.endTime, '%Y-%m-%d %H:%i:%s') endTime, 
       t2.round_teacher, t2.ter
FROM 
(
    -- 用户打赏日志表
    SELECT *
    FROM zbj_gift
    WHERE 
    --          date(sendDate) = '2024-12-30'
          date(sendDate) >= '2024-01-01' AND date(sendDate) <= '2024-12-31'
          AND terraceId IN(1, 3)
) t1
LEFT JOIN
(
    SELECT  CONCAT(round_date,' ',round_beginTime) beginTime,
           CONCAT(round_date,' ',CASE WHEN round_beginTime > round_endTime 
                                 THEN '23:59:59'
                                 ELSE round_endTime END) endTime,
           round_teacher,
           terraceId AS ter
    FROM zbj_round
    WHERE date(round_date) >= '2024-01-01' AND date(round_date) <= '2024-12-31'
          AND terraceId IN(1, 3)
) t2
ON (t1.sendDate >= t2.beginTime AND t1.sendDate <= t2.endTime 
    --    and t1.teacher= t2.round_teacher   
       AND t2.round_teacher LIKE CONCAT('%', t1.teacher, '%') 
       AND t1.terraceId = t2.ter
)
order by terraceId,sendDate;

难点&方法论沉淀

-- 【测试】检查 '浩柠老师/主播依然' 是否包含 '浩柠老师'
 SELECT '浩柠老师/主播依然' LIKE CONCAT('%','浩柠老师', '%');

-- 应用
 t2.round_teacher LIKE CONCAT('%', t1.teacher, '%')
 t2.round_teacher REGEXP t1.teacher

注意事项：性能影响：使用 REGEXP 通常会比 LIKE 产生更大的性能开销，特别是在处理大数据集时。因为 REGEXP 需要更复杂的模式匹配，而 MySQL 对 REGEXP 的优化能力相对较弱，可能无法像 LIKE 那样利用索引进行优化。