Id | FK_ID| Reference |
-----------------------
1    2100   GI2,GI32
2    2344   GI56

Id | FK_ID| Reference |
-----------------------
1    2100   GI2 
2    2100   GI32
3    2344   GI56

解决方法@H_301_12@

你真的不应该存储这样的数据.幸运的是,有一种方法可以通过递归sql来消除损坏,具体如下：

WITH unpivot (lvl,id,fk_ref,reference,tail) AS (  
  SELECT 1,CASE WHEN LOCATE(',',reference) > 0 
              THEN TRIM(LEFT(reference,LOCATE(',reference)-1))
              ELSE TRIM(reference) 
         END,reference) > 0 
              THEN SUBSTR(reference,reference)+1)    
              ELSE '' 
         END  
  FROM yourtable  
  UNION ALL  
  SELECT lvl + 1,tail) > 0 
              THEN TRIM(LEFT(tail,tail)-1))    
              ELSE TRIM(tail) 
         END,tail) > 0 
              THEN SUBSTR(tail,tail)+1)    
              ELSE '' 
         END
  FROM unpivot 
  WHERE lvl < 100 AND tail != '')
  SELECT id,reference FROM unpivot

PS.未经测试.