oracle 普通业务数据统计sql

-- 字符串分割示例
SELECT REGEXP_SUBSTR(‘17,20,23‘, ‘[^,]+‘, 1, LEVEL, ‘i‘) AS STR
  FROM DUAL  CONNECT BY LEVEL <=  LENGTH(‘17,20,23‘) - LENGTH(REGEXP_REPLACE(‘17,20,23‘, ‘,‘, ‘‘))+1;

CREATE TABLE tm_change(
     ID NUMBER,
     transit_zno VARCHAR2(10),
     src VARCHAR2(100),
     des VARCHAR2(100)
);

CREATE TABLE temp_tm_change(
     ID NUMBER,
     transit_zno VARCHAR2(10),
     src VARCHAR2(100),
     des VARCHAR2(100)
);

-- 原数据
SELECT t.*,ROWID FROM tm_change t;
--输出数据
SELECT t.* FROM temp_tm_change t;

DELETE FROM temp_tm_change;
COMMIT;

DECLARE
  recordCount NUMBER;-- 一条记录衍生的记录条数
  sumCount NUMBER; --总条数
BEGIN
  sumCount := 0;
  -- sum
  FOR rs IN (SELECT id,transit_zno ,src,des FROM tm_change) LOOP
      recordCount := 0;
       --src
       FOR srcRS IN (SELECT REGEXP_SUBSTR(rs.src, ‘[^,]+‘, 1, LEVEL, ‘i‘) AS str
           FROM DUAL  CONNECT BY LEVEL <=  LENGTH(rs.src) - LENGTH(REGEXP_REPLACE(rs.src, ‘,‘, ‘‘))+1) LOOP
        -- des
         FOR desRS IN (SELECT REGEXP_SUBSTR(rs.des, ‘[^,]+‘, 1, LEVEL, ‘i‘) AS str
             FROM DUAL  CONNECT BY LEVEL <=  LENGTH(rs.des) - LENGTH(REGEXP_REPLACE(rs.des, ‘,‘, ‘‘))+1) LOOP
             recordCount := recordCount + 1;
             sumCount := sumCount + 1;
             INSERT INTO temp_tm_change(ID,transit_zno,src,des) VALUES (rs.id,rs.transit_zno,srcRs.str,desRS.str);
             COMMIT;
         END LOOP;
       END LOOP;
       --dbms_output.put_line(‘insert a record : id ->‘ || rs.id || ‘  recordCount --> ‘ || recordCount ); buffer overflow
  END LOOP;
  --dbms_output.put_line(‘create  record  sumCount --> ‘ || sumCount );
END;

SELECT transit_zno,src,des FROM temp_tm_change GROUP BY transit_zno,src,des HAVING COUNT(*) > 1 ;