ORACLE大数据量下的分页解决方法: http://ddkangfu.blog.51cto.com/311989/1425797/
oracle 分页数据重复数据不正确: https://my.oschina.net/u/852445/blog/402325
实例：
SELECT *
FROM
(
SELECT RS.AMOUNT,RS.AREA,RS.AREANAME,[color=red]ROWNUM rn
FROM
(
SELECT (
CASE
WHEN oy.AMOUNT IS NULL THEN
ny.AMOUNT-0
WHEN ny.AMOUNT IS NULL THEN
0-oy.AMOUNT
ELSE
ny.AMOUNT-oy.AMOUNT
END) AS amount,NY.AREA,NY.areaName
FROM
(SELECT "SUM"(VI.AMOUNT) AS AMOUNT ,VI.AREA, dm.AREA_NAME as areaName FROM VI_INCOME_ITEM_MAPPING VI
LEFT JOIN DM_AREA dm ON VI.AREA = dm.AREA_CODE
WHERE VI.T_ITEM = #{BudgetTItem,jdbcType=VARCHAR} AND VI.BLEVEL ='3'
AND "SUBSTR"(VI.TXN_DATE,1,6) = ]]> #{startMonth,jdbcType=VARCHAR}
AND "SUBSTR"(VI.TXN_DATE,1,6) #{endMonth,jdbcType=VARCHAR}
GROUP BY VI.AREA,dm.AREA_NAME) ny
LEFT JOIN
(SELECT "SUM"(VI.AMOUNT) AS AMOUNT ,VI.AREA, dm.AREA_NAME as areaName FROM VI_INCOME_ITEM_MAPPING VI
LEFT JOIN DM_AREA dm ON VI.AREA = dm.AREA_CODE
WHERE VI.T_ITEM = #{BudgetTItem,jdbcType=VARCHAR} AND VI.BLEVEL ='3'
AND "SUBSTR"(VI.TXN_DATE,1,6) = ]]> #{lastStartMonth,jdbcType=VARCHAR}
AND "SUBSTR"(VI.TXN_DATE,1,6) #{lastEndMonth,jdbcType=VARCHAR}
GROUP BY VI.AREA,dm.AREA_NAME) oy
ON NY.AREA = OY.AREA
ORDER BY amount DESC ) [color=red]rs WHERE ROWNUM <= 10) WHERE rn >0
注意红色字体内层用ROWNUM，外层用RN，Oracle 一般第一次select时，有ROWNUM，可在第二层分页
分页两种方法：
第一种：
SELECT * FROM (SELECT t.*,ROWNUM as rowno FROM (SELECT * FROM test ORDER BY ADD_TIME ) t ) WHERE rowno BETWEEN 200001 AND 300000;
第二种：
SELECT * FROM ( SELECT t.*,ROWNUM as rowno FROM (SELECT * FROM test ORDER BY ADD_TIME) t WHERE ROWNUM <= 300000) WHERE rowno >=200001
第二种比第一种有效，因为第二种锁定了分页结果集，而第一种是从表所有记录中，分页；
对于两层嵌套的分页查询如：
实例：
select （select * from test ORDER BY ADD_TIME） where ROWNUM between 100000 and 200000
对于含排序（ORDER BY ADD_TIME）， 则排序字段ADD_TIME字段必须是唯一索引，否则已出现
结果集为空，或数据重复，含排序的正确做法如上面两种分页方式，再加一层嵌套。