ROWNUM是一种伪列，它会根据返回记录生成一个序列化的数字。利用ROWNUM，我们可以生产一些原先难以实现的结果输出，但因为它是伪列的这个特殊性，我们在使用时也需要注意一些事项，不要掉入“陷阱”。下面就介绍一下它的使用技巧及注意事项。
1 特殊结果输出
利用ROWNUM，我们可以做到一些特殊方式的输出。
1.1 Top N结果输出
我们如果希望取输出结果的前面几条数据，通过ROWNUM可以轻松实现：
sql> select * from t_test4 2 where rownum <= 5; USERNAME USER_ID CREATED------------------------------ ---------- ---------WOW 71 26-APR-07CS2 70 15-JAN-073 69 01-NOV-06DMP 68 12-OCT-06PROFILER 67 05-SEP-06但是，如果你希望对一个排序结果取Top N数据的话，使用ROWNUM存在一些“陷阱”，我们后面部分会介绍这些“陷阱”并且说明如何避免。
1.2 分页查询
利用ROWNUM对结果进行分页，下面返回结果中的第6到第10条记录：
sql> select * from 2 ( 3 select a.*, rownum as rn from css_bl_view a 4 where capture_phone_num = '(1) 925-4604800' 5 ) b 6 where b.rn between 6 and 10; 6 rows selected. Execution Plan---------------------------------------------------------- 0 SELECT STATEMENT Optimizer=CHOOSE (Cost=2770 Card=2183 Bytes =7166789) 1 0 VIEW (Cost=2770 Card=2183 Bytes=7166789) 2 1 COUNT 3 2 table ACCESS (FULL) OF 'CSS_BL_VIEW' (Cost=2770 Card=2183 Bytes=1305434) Statistics---------------------------------------------------------- 0 recursive calls 0 db block gets 29346 consistent gets 29190 physical reads 0 redo size 7328 bytes sent via sql*Net to client 234 bytes received via sql*Net from client 4 sql*Net roundtrips to/from client 0 sorts (memory) 0 sorts (disk) 5 rows processed另外一种实现方式：
sql> select * from css_bl_view a 2 where capture_phone_num = '(1) 925-4604800' 3 and rownum <= 10 4 minus 5 select * from css_bl_view a 6 where capture_phone_num = '(1) 925-4604800' 7 and rownum <= 5 8 ; Execution Plan---------------------------------------------------------- 0 SELECT STATEMENT Optimizer=CHOOSE (Cost=5920 Card=10 Bytes=8970) 1 0 MINUS 2 1 SORT (UNIQUE) (Cost=2960 Card=10 Bytes=5980) 3 2 COUNT (STOPKEY) 4 3 TABLE ACCESS (FULL) OF 'CSS_BL_VIEW' (Cost=2770 Card=2183 Bytes=1305434) 5 1 SORT (UNIQUE) (Cost=2960 Card=5 Bytes=2990) 6 5 COUNT (STOPKEY) 7 6 table ACCESS (FULL) OF 'CSS_BL_VIEW' (Cost=2770 Card=2183 Bytes=1305434) Statistics---------------------------------------------------------- 0 recursive calls 0 db block gets 62 consistent gets 50 physical reads 0 redo size 7232 bytes sent via sql*Net to client 234 bytes received via sql*Net from client 4 sql*Net roundtrips to/from client 2 sorts (memory) 0 sorts (disk) 5 rows processed第三种实现方式：
sql> select * from 2 ( 3 select a.*, rownum as rn from css_bl_view a 4 where capture_phone_num = '(1) 925-4604800' 5 and rownum <= 10 6 ) b 7 where b.rn > 5; Execution Plan---------------------------------------------------------- 0 SELECT STATEMENT Optimizer=CHOOSE (Cost=2770 Card=10 Bytes=32830) 1 0 VIEW (Cost=2770 Card=10 Bytes=32830) 2 1 COUNT (STOPKEY) 3 2 table ACCESS (FULL) OF 'CSS_BL_VIEW' (Cost=2770 Card=2183 Bytes=1305434) Statistics---------------------------------------------------------- 0 recursive calls 0 db block gets 35 consistent gets 30 physical reads 0 redo size 7271 bytes sent via sql*Net to client 234 bytes received via sql*Net from client 4 sql*Net roundtrips to/from client 0 sorts (memory) 0 sorts (disk) 5 rows processed这里特地将三种实现方式的查询计划及统计数据打印出来，大家可以比较一下3中方式的性能。
1.3 利用ROWNUM做分组子排序
对于以下表T_TEST4的内容：
OWNER NAME------------------------------------------------------STRMADMIN STREAMS_QUEUEAPARKMAN JOB_QUEUESYS AQ$_AQ_SRVNTFN_TABLE_ESYS AQ$_KUPC$DATAPUMP_QUETAB_EAPARKMAN AQ$_JMS_TEXT_ESTRMADMIN AQ$_STREAMS_QUEUE_TABLE_ESYS AQ$_SCHEDULER$_EVENT_QTAB_E…如果我们希望结果按照OWNER进行分组后，再对每组中成员进行编号，结果类似如下：
OWNER NO NAME------------------------------------------------------APARKMAN 1 JOB_QUEUE 2 AQ$_JMS_TEXT_ESTRMADMIN 1 STREAMS_QUEUE 2 AQ$_STREAMS_QUEUE_TABLE_ESYS 1 AQ$_AQ_SRVNTFN_TABLE_E 2 AQ$_KUPC$DATAPUMP_QUETAB_E 3 AQ$_SCHEDULER$_EVENT_QTAB_E…在没有ROWNUM时要实现这样的功能会很复杂，但通过ROWNUM我们可以轻松实现：
sql> SELECT DECODE(ROWNUM-min_sno,0,a.owner,NULL) owner,DECODE(ROWNUM-min_sno,0,1,rownum+1-min_sno) sno, a.name 2 FROM (SELECT * 3 FROM t_test8 4 ORDER BY owner, name ) a, 5 (SELECT owner, MIN(rownum) min_sno 6 FROM( SELECT * 7 FROM t_test8 8 ORDER BY owner, name) 9 GROUP BY owner) b 10 WHERE a.owner=b.owner; OWNER SNO NAME------------------------------ ---------- ------------------------------APARKMAN 1 JOB_QUEUE 2 AQ$_JMS_TEXT_ESTRMADMIN 1 STREAMS_QUEUE 2 AQ$_STREAMS_QUEUE_TABLE_ESYS 1 AQ$_AQ_SRVNTFN_TABLE_E 2 AQ$_KUPC$DATAPUMP_QUETAB_E 3 AQ$_SCHEDULER$_EVENT_QTAB_E 4 AQ$_SCHEDULER$_JOBQTAB_E 5 AQ$_STREAMS_QUEUE_TABLE_E 6 AQ$_SYS$SERVICE_METRICS_TAB_E 7 AQ$_AQ_EVENT_TABLE_E 8 AQ$_AQ$_MEM_MC_E 9 AQ$_ALERT_QT_E 10 ALERT_QUE 11 AQ_EVENT_TABLE_Q 12 SYS$SERVICE_METRICS 13 STREAMS_QUEUE 14 SRVQUEUE 15 SCHEDULER$_JOBQ 16 SCHEDULER$_EVENT_QUEUE 17 AQ_SRVNTFN_TABLE_QSYSMAN 1 AQ$_MGMT_NOTIFY_QTABLE_E 2 MGMT_NOTIFY_Qsystem 1 DEF$_AQERROR 2 DEF$_AQCALL 3 AQ$_DEF$_AQERROR_E 4 AQ$_DEF$_AQCALL_EWMSYS 1 AQ$_WM$EVENT_QUEUE_TABLE_E 2 WM$EVENT_QUEUE 29 rows selected.2 性能
我们很多程序员在确认某个表中是否有相应数据时，喜欢加上ROWNUM=1，其思路就是只要存在一条数据就说明有相应数据，查询就可以直接返回了，这样就能提高性能了。但是在10G之前，使用ROWNUM=1是不能达到预期的性能效果的，而是需要通过<2或<=1作为过滤条件才能达到预期效果，看以下查询计划：
sql> select * from t_test1 2 where object_id <100 3 and rownum = 1; Execution Plan---------------------------------------------------------- 0 SELECT STATEMENT Optimizer=CHOOSE (Cost=37 Card=1 Bytes=86) 1 0 COUNT (STOPKEY) 2 1 table ACCESS (BY INDEX ROWID) OF 'T_TEST1' (Cost=37 Card=89 Bytes=7654) 3 2 INDEX (RANGE SCAN) OF 'T_TEST1_PK' (UNIQUE) (Cost=2 Card=89) Statistics---------------------------------------------------------- 0 recursive calls 0 db block gets 62 consistent gets 0 physical reads 0 redo size 654 bytes sent via sql*Net to client 234 bytes received via sql*Net from client 4 sql*Net roundtrips to/from client 0 sorts (memory) 0 sorts (disk) 1 rows processed sql> select * from t_test1 2 where object_id <100 3 and rownum <= 1; Execution Plan---------------------------------------------------------- 0 SELECT STATEMENT Optimizer=CHOOSE (Cost=37 Card=1 Bytes=86) 1 0 COUNT (STOPKEY) 2 1 table ACCESS (BY INDEX ROWID) OF 'T_TEST1' (Cost=37 Card=89 Bytes=7654) 3 2 INDEX (RANGE SCAN) OF 'T_TEST1_PK' (UNIQUE) (Cost=2 Card=89) Statistics---------------------------------------------------------- 0 recursive calls 0 db block gets 3 consistent gets 0 physical reads 0 redo size 654 bytes sent via sql*Net to client 234 bytes received via sql*Net from client 4 sql*Net roundtrips to/from client 0 sorts (memory) 0 sorts (disk) 1 rows processed sql> / Execution Plan---------------------------------------------------------- 0 SELECT STATEMENT Optimizer=CHOOSE (Cost=37 Card=1 Bytes=86) 1 0 COUNT (STOPKEY) 2 1 table ACCESS (BY INDEX ROWID) OF 'T_TEST1' (Cost=37 Card=89 Bytes=7654) 3 2 INDEX (RANGE SCAN) OF 'T_TEST1_PK' (UNIQUE) (Cost=2 Card=89) Statistics---------------------------------------------------------- 0 recursive calls 0 db block gets 3 consistent gets 0 physical reads 0 redo size 654 bytes sent via sql*Net to client 234 bytes received via sql*Net from client 4 sql*Net roundtrips to/from client 0 sorts (memory) 0 sorts (disk) 1 rows processed10G以后，这个问题就被修正了：
sql> select * from t_test1 2 where rownum = 1; Execution Plan----------------------------------------------------------Plan hash value: 536364188 ------------------------------------------------------------------------------| Id | Operation | Name | Rows | Bytes | Cost (%CPU)| Time |------------------------------------------------------------------------------| 0 | SELECT STATEMENT | | 1 | 86 | 2 (0)| 00:00:01 ||* 1 | COUNT STOPKEY | | | | | || 2 | table ACCESS FULL| T_TEST1 | 1 | 86 | 2 (0)| 00:00:01 |------------------------------------------------------------------------------ Predicate Information (identified by operation id):--------------------------------------------------- 1 - filter(ROWNUM=1) Statistics---------------------------------------------------------- 1 recursive calls 0 db block gets 4 consistent gets 1 physical reads 0 redo size 1201 bytes sent via sql*Net to client 385 bytes received via sql*Net from client 2 sql*Net roundtrips to/from client 0 sorts (memory) 0 sorts (disk) 1 rows processed sql> select * from t_test1 2 where rownum <= 1; Execution Plan----------------------------------------------------------Plan hash value: 536364188 ------------------------------------------------------------------------------| Id | Operation | Name | Rows | Bytes | Cost (%CPU)| Time |------------------------------------------------------------------------------| 0 | SELECT STATEMENT | | 1 | 86 | 2 (0)| 00:00:01 ||* 1 | COUNT STOPKEY | | | | | || 2 | table ACCESS FULL| T_TEST1 | 1 | 86 | 2 (0)| 00:00:01 |------------------------------------------------------------------------------ Predicate Information (identified by operation id):--------------------------------------------------- 1 - filter(ROWNUM<=1) Statistics---------------------------------------------------------- 0 recursive calls 0 db block gets 4 consistent gets 0 physical reads 0 redo size 1201 bytes sent via sql*Net to client 385 bytes received via sql*Net from client 2 sql*Net roundtrips to/from client 0 sorts (memory) 0 sorts (disk) 1 rows processed3 ROWNUM的使用“陷阱”
由于ROWNUM是一个伪列，只有有结果记录时，ROWNUM才有相应数据，因此对它的使用不能向普通列那样使用，否则就会陷入一些“陷阱”当中。
3.1 对ROWNUM进行>、>=、=操作
不能对ROWNUM使用>（大于1的数值）、>=（大于或等于1的数值）、=（大于或等于1的数值），否则无结果
sql> select count(*) from css_bl_view a where rownum>0; COUNT(*)----------361928 sql> select count(*) from css_bl_view a 2 where rownum > 1; COUNT(*)---------- 0这是因为：
1、ROWNUM是伪列，必须要要有返回结果后，每条返回记录就会对应产生一个ROWNUM数值；
2、返回结果记录的ROWNUM是从1开始排序的，因此第一条始终是1;
这样，当查询到第一条记录时，该记录的ROWNUM为1，但条件要求ROWNUM>1，因此不符合，继续查询下一条；因为前面没有符合要求的记录，因此下一条记录过来后，其ROWNUM还是为1，如此循环，就不会产生结果。上述查询可以通过子查询来替代：
sql> select count(*) 2 from 3 (select BL_REF_CDE, rownum rn from css_bl_view) 4 where rn > 1; COUNT(*)---------- 361927我们可以通过以下方式来实现对ROWNUM的>、=的查询：
查询ROWNUM=5的数据：
sql> select object_id,object_name 2 from (select object_id,object_name, rownum as rn from t_test1) 3 where rn = 5; OBJECT_ID OBJECT_NAME---------- ------------------------------ 29 C_COBJ#查询ROWNUM > 25的数据：
sql> select * from t_test4 2 minus 3 select * from t_test4 4 where rownum <= 25; USERNAME USER_ID CREATED------------------------------ ---------- ---------DIP 19 21-NOV-05OUTLN 11 21-NOV-05PUBLIC 99999 18-JUL-07SYS 0 21-NOV-05SYSMAN 32 21-NOV-05system 5 21-NOV-05 6 rows selected.3.2 ROWNUM和Order BY
要注意的是：在使用ROWNUM时，只有当Order By的字段是主键时，查询结果才会先排序再计算ROWNUM，下面OBJECT_ID是表T_TEST1的主键字段：
sql> select object_id,object_name from t_test1 2 where rownum <= 5 3 order by object_id; OBJECT_ID OBJECT_NAME---------- ------------------------------ 2 C_OBJ# 3 I_OBJ# 4 TAB$ 5 CLU$ 6 C_TS#但是，对非主键字段OBJECT_NAME进行排序时，结果就混乱了：
sql> select object_id,object_name from t_test1 2 where rownum <= 5 3 order by object_name; OBJECT_ID OBJECT_NAME---------- ------------------------------ 28 CON$ 29 C_COBJ# 20 ICOL$ 44 I_USER1 15 UNDO$ sql> select count(*) from t_test1 2 where object_name < 'CON$'; COUNT(*)---------- 21645出现这种混乱的原因是：Oracle先按物理存储位置（rowid）顺序取出满足rownum条件的记录，即物理位置上的前5条数据，然后在对这些数据按照Order By的字段进行排序，而不是我们所期望的先排序、再取特定记录数。
如果需要对非主键字段排序再去前n条数据，我们可以以以下方式实现：
sql> select object_id,object_name 2 from (select object_id,object_name from t_test1 3 order by object_name) 4 where rownum <= 5; OBJECT_ID OBJECT_NAME---------- ------------------------------ 35489 /1000e8d1_LinkedHashMapValueIt 35490 /1000e8d1_LinkedHashMapValueIt 21801 /1005bd30_LnkdConstant 21802 /1005bd30_LnkdConstant 17205 /10076b23_OraCustomDatumClosur3.3 排序分页
当对存在重复值的字段排序后再分页输出，我们很容易会陷入到另外一个“陷阱”。
请看以下例子，我们希望对T_TEST1的OWNER字段排序后，以每页输出10个结果的方式分页输出：
sql> select owner, object_name from 2 (select a.*, rownum as rn from 3 (select owner, object_name from t_test1 order by owner) a 4 where rownum <= 10) 5 where rn >= 1; OWNER OBJECT_NAME------------------------------ ------------------------------AFWOWNER AFWADAPTERAFWOWNER AFWADAPTERCONFIGURATIONAFWOWNER AFWADAPTERCONFIGURATION_IDX1AFWOWNER AFWADAPTERFQN_PKAFWOWNER AFWADAPTERCONFIGURATION_PKAFWOWNER AFWADAPTERCONFIGURATION_IDX2AFWOWNER AFWSERVERCODE_PKAFWOWNER AFWSERVERAFWOWNER AFWADAPTERLOOKUP_IDX1AFWOWNER AFWADAPTERLOOKUP 10 rows selected. sql> select owner, object_name from 2 (select a.*, rownum as rn from 3 (select owner, object_name from t_test1 order by owner) a 4 where rownum <= 20) 5 where rn >= 11; OWNER OBJECT_NAME------------------------------ ------------------------------AFWOWNER AFWTOKENSTATUSCODE_PKAFWOWNER AFWTOKENSTATUSAFWOWNER AFWTOKENADMIN_IDX1AFWOWNER AFWTOKENADMINCODE_PKAFWOWNER AFWTOKENADMINAFWOWNER AFWTOKENAFWOWNER AFWSERVERCONFIGURATION_PKAFWOWNER AFWSERVERCONFIGURATIONAFWOWNER AFWSERVERAFWOWNER AFWADAPTERLOOKUP 10 rows selected.仔细比较结果，你会发现“AFWSERVER”、“AFWADAPTERLOOKUP”在两次分页结果中都出现了。但是OBJECT_NAME在每个OWNER中的值是唯一的，说明这个输出结果是错误的，我们又陷入了一个“陷阱”。这是怎么回事呢，请先看下上述语句的查询计划：
sql> select owner, object_name from 2 (select a.*, rownum as rn from 3 (select owner, object_name from t_test1 order by owner) a 4 where rownum <= 20) 5 where rn >= 11; 10 rows selected. Execution Plan---------------------------------------------------------- 0 SELECT STATEMENT Optimizer=CHOOSE (Cost=205 Card=20 Bytes=94 0) 1 0 VIEW (Cost=205 Card=20 Bytes=940) 2 1 COUNT (STOPKEY) 3 2 VIEW (Cost=205 Card=30670 Bytes=1042780) 4 3 SORT (ORDER BY STOPKEY) (Cost=205 Card=30670 Bytes=858760) 5 4 table ACCESS (FULL) OF 'T_TEST1' (Cost=42 Card=30670 Bytes=858760)看下这个“陷阱”是怎么形成的。从查询计划上，我们可以注意到，对于子查询
select a.*, rownum as rn from (select owner, object_name from t_test1 order by owner) awhere rownum <= 20优化器采用了“SORT (ORDER BY STOPKEY)”。
“SORT (ORDER BY STOPKEY)”不需要对所有数据进行排序，而是只要找出结果集中的按特定顺序的最前N条记录，一旦找出了这N条记录，就无需再对剩下的数据进行排序，而直接返回结果。这种算法我们可以视为是“快速排序”算法的变种。快速排序算法的基本思想是：先将数据分2组集合，保证第一集合中的每个数据都大于第二个集合中每个数据，然后再按这个原则对每个集合进行递归分组，直到集合的单位最小。在进行“SORT (ORDER BY STOPKEY)”时，首先找出N条数据（这些数据并没有做排序）放在第一组，保证第一组的数据都大于第二组的数据，然后只对第一组数据进行递归。
可以看到，基于这样的算法基础上，如果N的数值不同，数据的分组也不同（如N=20时，第一次分组比例为12:8，然后继续递归；当N=10时，第一次分组比例为3:7 … …），这样，在数据的排序字段值都相等时，输出结果的顺序就会因为N值不同而不同。
知道原因后，我们可以通过以下几种方法来避免这个“陷阱”。
1、让查询计划避免“SORT (ORDER BY STOPKEY)”，采用“SORT (ORDER BY)”，使数据排序不受ROWNUM的影响。但这样会使所有数据都做排序：
sql> select owner, object_name from 2 (select a.*, rownum as rn from 3 (select owner, object_name, rowid from t_test1 order by owner) a) 4 where rn <= 10 5 and rn >= 1; OWNER OBJECT_NAME------------------------------ ------------------------------AFWOWNER AFWADAPTERAFWOWNER AFWADAPTERCONFIGURATIONAFWOWNER AFWADAPTERCONFIGURATION_IDX2AFWOWNER AFWADAPTERCONFIGURATION_PKAFWOWNER AFWADAPTERCONFIGURATION_IDX1AFWOWNER AFWADAPTERFQN_PKAFWOWNER AFWADAPTERLOOKUP_IDX1AFWOWNER AFWSERVERCODE_PKAFWOWNER AFWSERVERCONFIGURATION_IDX1AFWOWNER AFWTOKENTYPECODE_PK 10 rows selected. sql> select owner, object_name from 2 (select a.*, rownum as rn from 3 (select owner, object_name, rowid from t_test1 order by owner) a) 4 where rn <= 20 5 and rn >= 11; OWNER OBJECT_NAME------------------------------ ------------------------------AFWOWNER AFWTOKENTYPEAFWOWNER AFWTOKENSTATUSCODE_PKAFWOWNER AFWTOKENSTATUSAFWOWNER AFWTOKENADMIN_IDX1AFWOWNER AFWTOKENADMINCODE_PKAFWOWNER AFWTOKENADMINAFWOWNER AFWTOKENAFWOWNER AFWSERVERCONFIGURATION_PKAFWOWNER AFWTOKEN_PKAFWOWNER AFWTOKEN_IDX6 10 rows selected. sql> set autot tracesql> select owner, object_name from 2 (select a.*, rownum as rn from 3 (select owner, object_name, rowid from t_test1 order by owner) a) 4 where rn <= 20 5 and rn >= 11; 10 rows selected. Execution Plan---------------------------------------------------------- 0 SELECT STATEMENT Optimizer=CHOOSE (Cost=237 Card=30670 Bytes=1441490) 1 0 VIEW (Cost=237 Card=30670 Bytes=1441490) 2 1 COUNT 3 2 VIEW (Cost=237 Card=30670 Bytes=1042780) 4 3 SORT (ORDER BY) (Cost=237 Card=30670 Bytes=1073450) 5 4 table ACCESS (FULL) OF 'T_TEST1' (Cost=42 Card=30670 Bytes=1073450)2、在排序时，加上一个或多个字段（如主键字段、ROWID），使排序结果具有唯一性：
sql> select owner, object_name from 2 (select a.*, rownum as rn from 3 (select owner, object_name, rowid from t_test1 order by owner, object_id) a 4 where rownum <= 10) 5 where rn >= 1; OWNER OBJECT_NAME------------------------------ ------------------------------AFWOWNER AFWADAPTERAFWOWNER AFWADAPTERFQN_PKAFWOWNER AFWADAPTERCONFIGURATIONAFWOWNER AFWADAPTERCONFIGURATION_PKAFWOWNER AFWADAPTERCONFIGURATION_IDX1AFWOWNER AFWADAPTERCONFIGURATION_IDX2AFWOWNER AFWADAPTERLOOKUPAFWOWNER AFWADAPTERLOOKUP_IDX1AFWOWNER AFWSERVERAFWOWNER AFWSERVERCODE_PK 10 rows selected. sql> select owner, object_name from 2 (select a.*, rownum as rn from 3 (select owner, object_name, rowid from t_test1 order by owner, object_id) a 4 where rownum <= 20) 5 where rn >= 11; OWNER OBJECT_NAME------------------------------ ------------------------------AFWOWNER AFWSERVERCONFIGURATIONAFWOWNER AFWSERVERCONFIGURATION_PKAFWOWNER AFWSERVERCONFIGURATION_IDX1AFWOWNER AFWTOKENAFWOWNER AFWTOKEN_PKAFWOWNER AFWTOKEN_IDX1AFWOWNER AFWTOKEN_IDX2AFWOWNER AFWTOKEN_IDX3AFWOWNER AFWTOKEN_IDX4AFWOWNER AFWTOKEN_IDX5 10 rows selected. sql> set autot tracesql> select owner, object_name from 2 (select a.*, rownum as rn from 3 (select owner, object_name, rowid from t_test1 order by owner, object_id) a 4 where rownum <= 20) 5 where rn >= 11; 10 rows selected. Execution Plan---------------------------------------------------------- 0 SELECT STATEMENT Optimizer=CHOOSE (Cost=253 Card=20 Bytes=940) 1 0 VIEW (Cost=253 Card=20 Bytes=940) 2 1 COUNT (STOPKEY) 3 2 VIEW (Cost=253 Card=30670 Bytes=1042780) 4 3 SORT (ORDER BY STOPKEY) (Cost=253 Card=30670 Bytes=1196130) 5 4 table ACCESS (FULL) OF 'T_TEST1' (Cost=42 Card=306 70 Bytes=1196130)3、对排序字段建立索引，并强制使用索引。这样就能利用索引已经建立好的排序结果：
sql> create index t_test1_idx1 on t_test1(owner); Index created. sql> select owner, object_name from 2 (select a.*, rownum as rn from 3 (select /*+index(t T_TEST1_IDX1)*/owner, object_name from t_test1 t order by owner) a 4 where rownum <= 10) 5 where rn >= 1 6 ; OWNER OBJECT_NAME------------------------------ ------------------------------AFWOWNER AFWADAPTERAFWOWNER AFWADAPTERCONFIGURATIONAFWOWNER AFWADAPTERCONFIGURATION_IDX1AFWOWNER AFWADAPTERCONFIGURATION_IDX2AFWOWNER AFWADAPTERCONFIGURATION_PKAFWOWNER AFWADAPTERFQN_PKAFWOWNER AFWADAPTERLOOKUPAFWOWNER AFWADAPTERLOOKUP_IDX1AFWOWNER AFWSERVERAFWOWNER AFWSERVERCODE_PK 10 rows selected. sql> select owner, object_name from 2 (select a.*, rownum as rn from 3 (select /*+index(t T_TEST1_IDX1)*/owner, object_name from t_test1 t order by owner) a 4 where rownum <= 20) 5 where rn >= 11; OWNER OBJECT_NAME------------------------------ ------------------------------AFWOWNER AFWSERVERCONFIGURATIONAFWOWNER AFWSERVERCONFIGURATION_IDX1AFWOWNER AFWSERVERCONFIGURATION_PKAFWOWNER AFWTOKENAFWOWNER AFWTOKENADMINAFWOWNER AFWTOKENADMINCODE_PKAFWOWNER AFWTOKENADMIN_IDX1AFWOWNER AFWTOKENSTATUSAFWOWNER AFWTOKENSTATUSCODE_PKAFWOWNER AFWTOKENTYPE 10 rows selected. sql> set autot tracesql> select owner, object_name from 2 (select a.*, rownum as rn from 3 (select /*+index(t T_TEST1_IDX1)*/owner, object_name from t_test1 t order by owner) a 4 where rownum <= 20) 5 where rn >= 11; 10 rows selected. Execution Plan---------------------------------------------------------- 0 SELECT STATEMENT Optimizer=CHOOSE (Cost=414 Card=20 Bytes=940) 1 0 VIEW (Cost=414 Card=20 Bytes=940) 2 1 COUNT (STOPKEY) 3 2 VIEW (Cost=414 Card=30670 Bytes=1042780) 4 3 table ACCESS (BY INDEX ROWID) OF 'T_TEST1' (Cost=414Card=30670 Bytes=858760) 5 4 INDEX (FULL SCAN) OF 'T_TEST1_IDX1' (NON-UNIQUE) ( Cost=26 Card=30670) 以上就是ROWNUM的使用技巧及其注意事项，希望编程成员正确使用ROWNUM，也希望DBA遇到相关问题能迅速定位。
转载至：http://blog.csdn.net/nestashen/article/details/3117729
Oracle中ROWNUM的使用技巧的更多相关文章
-
ORACLE 中ROWNUM用法总结(转)
ORACLE 中ROWNUM用法总结! 对于 Oracle 的 rownum 问题,很多资料都说不支持>,>=,=,between...and,只能用以上符号(<.<=.!=) ...
-
oracle 中 rownum 和 row_number()
简单的介绍下oracle 中rownum 和 row_number() 使用,实例演示. 参照:http://www.cnblogs.com/zjrstar/archive/2006/08/31/49 ...
-
oracle中rownum和rowid的区别
rownum和rowid的区别总括: rownum和rowid都是伪列,但是两者的根本是不同的. rownum是根据sql查询出的结果给每行分配一个逻辑编号,所以你的sql不同也就会导致最终rownu ...
-
ORACLE 中ROWNUM用法总结!
ORACLE 中ROWNUM用法总结! 对于 Oracle 的 rownum 问题,很多资料都说不支持>,>=,=,between...and,只能用以上符号(<.<=.!=) ...
-
Oracle中rownum的基本用法
Oracle中rownum的基本用法 对于rownum来说它是oracle系统顺序分配为从查询返回的行的编号,返回的第一行分配的是1,第二行是2,依此类推,这个伪字段可以用于限制查询返回的总行数,且r ...
-
ORACLE 中ROWNUM
ORACLE 中ROWNUM用法总结! 对于 Oracle 的 rownum 问题,很多资料都说不支持>,>=,=,between...and,只能用以上符号(<.<=.!=) ...
-
Oracle中rownum和rowid的理解
rownum,rowid都叫伪列. 但是,rownum是逻辑上的编号,且其值总是从1开始,每行的rounum不是固定的.而rowid是“物理”编号.若数据库文件没有移动,则每行的 rowid一般是固定 ...
-
oracle中rownum和rowid的区别【转】
rownum和rowid的区别总括: rownum和rowid都是伪列,但是两者的根本是不同的. rownum是根据sql查询出的结果给每行分配一个逻辑编号,所以你的sql不同也就会导致最终rownu ...
-
转：ORACLE 中ROWNUM用法总结!
oracle 分页查询语句:select * from (select u.*,rownum r from (select * from userifno) u where rownum<大值) ...
随机推荐
-
【代码笔记】iOS-多张图片合成一张
代码: RootViewController.m #import "RootViewController.h" @interface RootViewController () @ ...
-
webpack常用插件
extract-text-wepback-plugin 该插件用于把css代码从页面中抽离出来,以link的形式从外部加载 html-webpack-plugin 可以自动快速地生成html文件
-
C++11 智能指针unique_ptr使用 -- 以排序二叉树为例
用智能指针可以简化内存管理.以树为例,如果用普通指针,通常是在插入新节点时用new,在析构函数中调用delete:但有了unique_ptr类型的智能指针,就不需要在析构函数中delete了,因为当u ...
-
POJ 3264-Balanced Lineup, NYOJ 119-士兵杀敌3 线段树
士兵杀敌(三) 时间限制:2000 ms | 内存限制:65535 KB 难度:5 描述 南将军统率着N个士兵,士兵分别编号为1~N,南将军经常爱拿某一段编号内杀敌数最高的人与杀敌数最低的人进行比 ...
-
在JSP中使用BootStrap
1. 下载BootStrap,然后再Jsp标签中添加如下标签: <html> <head lang="zh-cn"> <meta charset=&q ...
-
Majority Element II——LeetCode
Given an integer array of size n, find all elements that appear more than ⌊ n/3 ⌋ times. The algorit ...
-
HDU_2017——统计数字字符出现的个数
Problem Description 对于给定的一个字符串,统计其中数字字符出现的次数. Input 输入数据有多行,第一行是一个整数n,表示测试实例的个数,后面跟着n行,每行包括一个由字母和数 ...
-
一个JAVA代码
public class HelloJava { public static void main(String[] args) { System.out.println("这"); ...
-
重构ConditionHelper
在上一篇日志写了ConditionHelper的具体实现方法和它的用法,但那时只是把心里所想的一个思路用代码记录下来,而没有经过实践检 验的,难免存在一些问题,但在最近的一个新项目中用到Conditi ...
-
Nginx memcached应用层反向代理
L:105 Module ngx_http_memcached_module 模块默认编译进Nginx