oracle分析函数系列之LAG和LEAD:取出同一字段的前N行的数据和后N行的值
另一篇博文有关于此函数的变通使用:http://blog.csdn.net/rfb0204421/article/details/7609520
Lag和Lead函数可以在一次查询中取出同一字段的前N行的数据和后N行的值。这种操作可以使用对相同表的表连接来实现，不过使用LAG和LEAD有更高的效率。以下是LAG和LEAD的例子：
SQL> select year,region,profit ,lag (profit,1) over (order by year)
2 as last_year_exp from test;
YEAR REGION PROFIT LAST_YEAR_EXP
---- ------- ---------- -------------
2003 West 88
2003 West 88 88
2003 Central 101 88
2003 Central 100 101
2003 East 102 100
2004 West 77 102
2004 East 103 77
2004 West 89 103
SQL> select year,region,profit ,lead (profit,1) over (order by year)
2 as next_year_exp from test;
YEAR REGION PROFIT NEXT_YEAR_EXP
---- ------- ---------- -------------
2003 West 88 88
2003 West 88 101
2003 Central 101 100
2003 Central 100 102
2003 East 102 77
2004 West 77 103
2004 East 103 89
2004 West 89
Lag函数为Lag(exp,N,defval)，defval是当该函数无值可用的情况下返回的值。Lead函数的用法类似。
Lead和Lag函数也可以使用分组，以下是使用region分组的例子：
SQL> select year,region,profit ,
2 lag (profit,1,0) over (PARTITION BY region order by year)
3 as last_year_exp from test;
YEAR REGION PROFIT LAST_YEAR_EXP
---- ------- ---------- -------------
2003 Central 101 0
2003 Central 100 101
2003 East 102 0
2004 East 103 102
2003 West 88 0
2003 West 88 88
2004 West 77 88
2004 West 89 77
该函数也可以用于纵向表转横向表、解释示例：
SQL> select * from test_value;
MONS JJR CJL CJJE
---------- ---------- ---------- ----------
200801 LZF 250 1999
200802 LZF 200 2000
200803 LZF 300 1000
200804 LZF 23 189
200805 LZF 356 456
200806 LZF 100 200
200807 LZF 600 700
200808 LZF 23 123
200809 LZF 400 500
9 rows selected
SQL>
SQL> select rownum 序号,Mons,cjl cjl_01,
2 lead(cjl,1) over (order by mons desc) cjl_02,
3 lead(cjl,2) over (order by mons desc) cjl_03,
4 lead(cjl,3) over (order by mons desc) cjl_04,
5 lead(cjl,4) over (order by mons desc) cjl_05,
6 lead(cjl,5) over (order by mons desc) cjl_06,
7 lead(cjl,6) over (order by mons desc) cjl_07,
8 lead(cjl,7) over (order by mons desc) cjl_08,
9 lead(cjl,8) over (order by mons desc) cjl_09
10 from test_value
SQL> select * from test_value;
MONS JJR CJL CJJE
---------- ---------- ---------- ----------
200801 LZF 250 1999
200802 LZF 200 2000
200803 LZF 300 1000
200804 LZF 23 189
200805 LZF 356 456
200806 LZF 100 200
200807 LZF 600 700
200808 LZF 23 123
200809 LZF 400 500
9 rows selected
SQL>
SQL> select rownum 序号,Mons,cjl cjl_01,
2 lead(cjl,1) over (order by mons desc) cjl_02,
3 lead(cjl,2) over (order by mons desc) cjl_03,
4 lead(cjl,3) over (order by mons desc) cjl_04,
5 lead(cjl,4) over (order by mons desc) cjl_05,
6 lead(cjl,5) over (order by mons desc) cjl_06,
7 lead(cjl,6) over (order by mons desc) cjl_07,
8 lead(cjl,7) over (order by mons desc) cjl_08,
9 lead(cjl,8) over (order by mons desc) cjl_09
10 from test_value
序号 MONS CJL_01 CJL_02 CJL_03 CJL_04 CJL_05 CJL_06 CJL_07 CJL_08 CJL_09
---------- ---------- ---------- ---------- ---------- ---------- ---------- ---------- ---------- ---------- ----------
9 200809 400 23 600 100 356 23 300 200 250
8 200808 23 600 100 356 23 300 200 250
7 200807 600 100 356 23 300 200 250
6 200806 100 356 23 300 200 250
5 200805 356 23 300 200 250
4 200804 23 300 200 250
3 200803 300 200 250
2 200802 200 250
1 200801 250
---------- ---------- ---------- ---------- ---------- ---------- ---------- ---------- ---------- ---------- ----------
9 200809 400 23 600 100 356 23 300 200 250
8 200808 23 600 100 356 23 300 200 250
7 200807 600 100 356 23 300 200 250
6 200806 100 356 23 300 200 250
5 200805 356 23 300 200 250
4 200804 23 300 200 250
3 200803 300 200 250
2 200802 200 250
1 200801 250