update

content/posts/continuous_time_segments.md

@@ -0,0 +1,57 @@
++++
+title = "连续时间片段判断解决方案"
+date = "2024-11-07"
++++
+
+## 场景
+一个账号同一个会议室不能连续多天预定(4天)超过4个小时(会议预定开始时间和结束时间不垮天)
+
+## schema
+```mysql
+create table meeting_room_reserve
+(
+    id              bigint auto_increment comment '主键'
+        primary key,
+    meeting_room_id bigint                   null comment '会议室id',
+    start_time      varchar(24) charset utf8 null comment '会议开始时间',
+    end_time        varchar(24) charset utf8 null comment '会议结束时间',
+    meeting_date    varchar(32) charset utf8 null comment '会议日期'
+);
+```
+
+## 思路
+1. 查出满足预定条件的会议记录（超过4小时、当前账号）
+2. 即将预定的日期union步骤1查询结果的日期按照预定时间排序
+3. 查询排序后的日期与前一个日期的时间间隔天数
+4. 设置group_id, 间隔天数为1的日期与前一天是同一组
+5. 根据group_id分组, 每组都是一个连续时间区间, 过滤出当前日期所在的分组
+6. 根据连续时间天数继续过滤，如果不满足要求返回结果是1，满足是0
+
+
+## final solution
+```mysql
+select count(*)
+        from (WITH DateList AS (SELECT #{meetingDate} AS meeting_date
+                                UNION
+                                SELECT meeting_date
+                                FROM t_meeting_pre_form t
+                                WHERE t.create_user = #{userCode}
+                                  AND TIMESTAMPDIFF(HOUR, STR_TO_DATE(t.meeting_time_top, '%Y-%m-%d %H:%i:%s'),
+                                                    STR_TO_DATE(t.meeting_time_bottom, '%Y-%m-%d %H:%i:%s')) > 4
+                                  and t.delete_mark = 0
+                                  and t.meeting_room_id = #{meetingRoomId}
+                                ORDER BY meeting_date),
+                   RankedDates AS (SELECT meeting_date,
+                                          DATEDIFF(meeting_date,
+                                                   LAG(meeting_date) OVER (ORDER BY meeting_date)) AS date_diff
+                                   FROM DateList),
+                   ConsecutiveGroups AS (SELECT meeting_date,
+                                                SUM(IF(date_diff = 1, 0, 1))
+                                                    OVER (ORDER BY meeting_date) AS group_id
+                                         FROM RankedDates)
+              SELECT group_id, MIN(meeting_date) AS start_date, MAX(meeting_date) AS end_date
+              FROM ConsecutiveGroups
+              GROUP BY group_id
+              HAVING DATEDIFF(MAX(meeting_date), MIN(meeting_date)) >= 3
+                 and str_to_date(#{meetingDate}, '%Y-%m-%d') between str_to_date(start_date, '%Y-%m-%d') and str_to_date(end_date, '%Y-%m-%d')) as tmp
+```

content/posts/dm-migrate.md

@@ -14,6 +14,8 @@ date = "2024-10-28"
 6. convert => cast
 7. convert(str, SIGNED) => to_number(REGEXP_SUBSTR(str, '[0-9]+'))
 8. match_partern = REPLACE('${ew.department}', ',', '|'), REGEXP  match_partern => REGEXP_LIKE(var1, match_partern)
+9. if函数传参数为null时（部分环境支持，兼容性不好），替换为case ... when ... else ...end
+10. DATEDIFF 必须指定date_unit
 
 # 数据迁移
 1. [迁移问题](https://eco.dameng.com/document/dm/zh-cn/faq/faq-mysql-dm8-migrate.html)
@@ -48,3 +50,13 @@ GROUP BY
     i.OWNER, i.TABLE_NAME, i.INDEX_NAME
 ORDER BY i.INDEX_NAME
 ```
+
+## 相关数据操作
+1. 清空数据库表
+```sql
+BEGIN
+    FOR rec IN (SELECT TABLE_NAME FROM ALL_TABLES WHERE OWNER = 'XZFW') LOOP
+        EXECUTE IMMEDIATE 'DROP TABLE XZFW.' || rec.TABLE_NAME || ' CASCADE CONSTRAINTS';
+    END LOOP;
+END;
+```

content/posts/quarts.md

@@ -0,0 +1,4 @@
++++
+title = "quarts in action"
+date = "2024-11-05"
++++