oracle值sql记录

前言

由于本人sql比较薄弱将一些常用又使用的sql语句记录与大家分享
一：替换特殊字符分割的字符串

需求如下：

有张原始表tschedule 和新表bjcodechange ，需要将原始表的compeny_code字段和bjcodechange 的oldwindcode 进行关联替换成bjcodechange表的newwindcode字段

无特殊字符的查询很简单如下

sql 复制代码

UPDATE tschedule  ts
   SET ts.security_code =
       (SELECT bj.newwindcode
          FROM BJcodechange bj
         WHERE ts.security_code = bj.oldwindcode
           )
 WHERE EXISTS (SELECT 1
          FROM BJcodechange bj
         WHERE ts.security_code = bj.oldwindcode);

但是有特殊字符的数据如下注意测试原始表的compeny_code字段的内容许如下所示

tschedule：

name	compeny_code
张三	000001
李四	800981,000001,000002

bjcodechange：

oldwindcode	newwindcode
800981	888888

最终的效果如下

name	compeny_code
张三	000001
李四	888888 ,000001,000002

完整步骤

步骤1：拆分后的中间数据 (split_data)

rid	elem	ord
OKABKyAAA2AAAAiTAAD	000001	1
OKABKyAAA2AAAAiTAAD	800981	2
OKABKyAAA2AAAAiTAAD	000001	3
AABKyAAA2AAAAiTAAD	000002	1
步骤2：关联后的中间数据
rid	elem	newwindcode
----------------------	--------	-------------
OKABKyAAA2AAAAiTAAD	430017	920017
OKABKyAAA2AAAAiTAAD	000001	(null)
OKABKyAAA2AAAAiTAAD	430017	920017
AABKyAAA2AAAAiTAAD	000001	(null)
步骤3：聚合后的结果 (merged)
rid	new_code
----------------------	---------------------
OKABKyAAA2AAAAiTAAD	920017,000001,920017
AABKyAAA2AAAAiTAAD	000001

sql 复制代码

MERGE INTO tschedule a
USING (
  WITH split_data AS (
    SELECT 
      a.rowid AS rid, 
      regexp_substr(a.compeny_code, '[^,]+', 1, LEVEL) AS elem,
      LEVEL AS ord
    FROM tschedule a
    CONNECT BY 
      LEVEL <= regexp_count(a.compeny_code, ',') + 1
      AND PRIOR a.rowid = a.rowid
      AND PRIOR DBMS_RANDOM.VALUE IS NOT NULL
  )
  SELECT 
    rid,
    LISTAGG(NVL(b.newwindcode, elem), ',') WITHIN GROUP (ORDER BY ord) AS new_code 
  FROM split_data sd
  LEFT JOIN bjcodechange b ON sd.elem = b.oldwindcode  
  GROUP BY rid
) merged
ON (a.rowid = merged.rid)
WHEN MATCHED THEN
  UPDATE SET a.compeny_code = merged.new_code
WHERE a.compeny_code != merged.new_code;

好了至此 oracle值sql记录点点关注不迷路老铁们！！！！！