全局（Global) 與本地（Local)索引的區別
一、定義說明
oracle數據庫中，存在多種對象，最常見的表和索引，索引的目的是為了加快查詢；oracle建議一個表>2g時，就建議進行分區，分區表的好處此處省略，分區表有維護操作，但是某些維護操作對整個表的索引維護造成巨大開銷，因此本地索引就是解決分區表管理的問題。
非分區索引創建方式，create index index_name on table_name(column_name);
查詢索引狀態從dba_indexes->status列查詢
全局分區索引創建，create index index_name on table_name(column_name) global partition by range(column_name) (partition p1 values less than (11104) tablespace system,partition p2 values less than (11100) tablespace users,partition p3 values less than (maxvalue) tablespace sysaux.) or partition by hash (id) partitions 8 online等;
查詢索引狀態dba_ind_partitions->STATUS列查詢
查詢分區索引類型，是全局or本地 DBA_PART_INDEXES ->LOCALITY
本地分區索引創建，create index index_name on table_name(column_name) local;
查詢索引狀態dba_ind_partitions->STATUS列查詢
二、create/selet 說明
1) 非分區索引 SQL> create index ind_1 on range_part_tab(deal_date); Index created. SQL> select index_name,PARTITION_NAME,HIGH_VALUE,STATUS,TABLESPACE_NAME from dba_ind_partitions where index_owner=‘SYS‘ and index_name=‘IND_1‘; no rows selected SQL> select index_name,STATUS,GLOBAL_STATS,TABLESPACE_NAME from dba_indexes where owner=‘SYS‘ and index_name=‘IND_1‘; INDEX_NAME STATUS GLO TABLESPACE_NAME------------------------------ -------- --- ------------------------------ IND_1 VALID YES SYSTEM STATUS 表示一個非分區索引是否VALID或UNUSABLE GLOBAL_STATUS 對於分區索引，指示是通過分析整個索引（YES）來收集統計信息，還是根據基礎索引分區和子分區（NO）的統計信息進行估計
2）全局分區索引
create index ind_3 on range_part_tab(area_code) global partition by range(area_code) (partition p1 values less than (11104) tablespace system,partition p2 values less than (11100) tablespace users,partition p3 values less than (maxvalue) tablespace sysaux)
*
ERROR at line 1:
ORA-14037: partition bound of partition "P1" is too high
12:39:59 SYS@ceshi>select area_code,count(*) from range_part_tab group by area_code;
AREA_CODE COUNT(*)
---------- ----------
599 11104
594 11074
596 11021
593 11113
595 10927
592 11214
598 11235
591 11169
597 11143
9 rows selected. 全局範圍分區索引，第一個分區不能比所有的值都大
SYS@ceshi>create index ind_3 on range_part_tab(area_code) global partition by range(area_code) (partition p1 values less than (592) tablespace system,partition p2 values less than (11100) tablespace users,partition p3 values less than (maxvalue) tablespace sysaux);
Index created.
select index_name,PARTITION_NAME,HIGH_VALUE,STATUS,TABLESPACE_NAME from dba_ind_partitions where index_owner=‘SYS‘ and index_name=‘IND_3‘
INDEX_NAME PARTITION_ HIGH_VALUE STATUS TABLESPACE_NAME
---------- ---------- ------------------------------ -------- ------------------------------
IND_3 P1 592 USABLE SYSTEM
IND_3 P2 11100 USABLE USERS
IND_3 P3 MAXVALUE USABLE SYSAUX
PARTITION_NAME 索引的分區名稱
HIGH_VALUE 索引分區的範圍條件
STATUS 分區索引的狀態
SYS@ceshi>select index_name,table_name,PARTITIONING_TYPE,SUBPARTITIONING_TYPE,PARTITION_COUNT,LOCALITY,DEF_TABLESPACE_NAME from DBA_PART_INDEXES where index_name=‘IND_3‘;
INDEX_NAME TABLE_NAME PARTITION SUBPARTIT PARTITION_COUNT LOCALI DEF_TABLESPACE_NAME
---------- ------------------------------ --------- --------- --------------- ------ ------------------------------
IND_3 RANGE_PART_TAB RANGE NONE 3 GLOBAL SYSTEM
hash全局分區索引，使用場景較多，例如索引塊爭用，且表為非分區表時
SYS@ceshi>create table a1 as select * from range_part_tab;
SYS@ceshi>create index ind_4 on a1(id) global partition by hash (id) partitions 8 online;
Index created.
SYS@ceshi>alter index ind_4 parallel 1;
Index altered.
select index_name,PARTITION_NAME,HIGH_VALUE,STATUS,TABLESPACE_NAME from dba_ind_partitions where index_owner=‘SYS‘ and index_name=‘IND_4‘
INDEX_NAME PARTITION_NAME HIGH_VALUE STATUS TABLESPACE_NAME
------------------------------ ------------------------------ ---------- -------- ------------------------------
IND_4 SYS_P61 USABLE SYSTEM
IND_4 SYS_P62 USABLE SYSTEM
IND_4 SYS_P63 USABLE SYSTEM
IND_4 SYS_P64 USABLE SYSTEM
IND_4 SYS_P65 USABLE SYSTEM
IND_4 SYS_P66 USABLE SYSTEM
IND_4 SYS_P67 USABLE SYSTEM
IND_4 SYS_P68 USABLE SYSTEM
8 rows selected.
SYS@ceshi>select index_name,table_name,PARTITIONING_TYPE,SUBPARTITIONING_TYPE,PARTITION_COUNT,LOCALITY,DEF_TABLESPACE_NAME from DBA_PART_INDEXES where index_name=‘IND_4‘;
INDEX_NAME TABLE_NAME PARTITION SUBPARTIT PARTITION_COUNT LOCALI DEF_TABLESPACE_NAME
------------------------------ ------------------------------ --------- --------- --------------- ------ ------------------------------
IND_4 A1 HASH NONE 8 GLOBAL SYSTEM
3） 本地分區索引
SQL> create index ind_2 on range_part_tab(id) local;
Index created.
select index_name,PARTITION_NAME,HIGH_VALUE,STATUS,TABLESPACE_NAME from dba_ind_partitions where index_owner=‘SYS‘ and index_name=‘IND_2‘
INDEX_NAME PARTITION_ HIGH_VALUE STATUS TABLESPACE_NAME
------------------ -----------------------------------------------------------------------------------------------------------------------
IND_2 P1 TO_DATE(‘ 2017-02-01 00:00:00‘, ‘SYYYY-MM-DD HH24:MI:SS‘, ‘NLS_CALENDAR=GREGORIAN‘) USABLE SYSTEM
IND_2 P10 TO_DATE(‘ 2017-11-01 00:00:00‘, ‘SYYYY-MM-DD HH24:MI:SS‘, ‘NLS_CALENDAR=GREGORIAN‘) USABLE SYSTEM
IND_2 P11 TO_DATE(‘ 2017-12-01 00:00:00‘, ‘SYYYY-MM-DD HH24:MI:SS‘, ‘NLS_CALENDAR=GREGORIAN‘) USABLE SYSTEM
IND_2 P12 TO_DATE(‘ 2018-01-01 00:00:00‘, ‘SYYYY-MM-DD HH24:MI:SS‘, ‘NLS_CALENDAR=GREGORIAN‘) USABLE SYSTEM
IND_2 P2 TO_DATE(‘ 2017-03-01 00:00:00‘, ‘SYYYY-MM-DD HH24:MI:SS‘, ‘NLS_CALENDAR=GREGORIAN‘) USABLE SYSTEM
IND_2 P3 TO_DATE(‘ 2017-04-01 00:00:00‘, ‘SYYYY-MM-DD HH24:MI:SS‘, ‘NLS_CALENDAR=GREGORIAN‘) USABLE SYSTEM
IND_2 P4 TO_DATE(‘ 2017-05-01 00:00:00‘, ‘SYYYY-MM-DD HH24:MI:SS‘, ‘NLS_CALENDAR=GREGORIAN‘) USABLE SYSTEM
IND_2 P5 TO_DATE(‘ 2017-06-01 00:00:00‘, ‘SYYYY-MM-DD HH24:MI:SS‘, ‘NLS_CALENDAR=GREGORIAN‘) USABLE SYSTEM
IND_2 P6 TO_DATE(‘ 2017-07-01 00:00:00‘, ‘SYYYY-MM-DD HH24:MI:SS‘, ‘NLS_CALENDAR=GREGORIAN‘) USABLE SYSTEM
IND_2 P7 TO_DATE(‘ 2017-08-01 00:00:00‘, ‘SYYYY-MM-DD HH24:MI:SS‘, ‘NLS_CALENDAR=GREGORIAN‘) USABLE SYSTEM
IND_2 P8 TO_DATE(‘ 2017-09-01 00:00:00‘, ‘SYYYY-MM-DD HH24:MI:SS‘, ‘NLS_CALENDAR=GREGORIAN‘) USABLE SYSTEM
IND_2 P9 TO_DATE(‘ 2017-10-01 00:00:00‘, ‘SYYYY-MM-DD HH24:MI:SS‘, ‘NLS_CALENDAR=GREGORIAN‘) USABLE SYSTEM
IND_2 P_MAX MAXVALUE USABLE SYSTEM
13 rows selected.
SYS@ceshi>select index_name,table_name,PARTITIONING_TYPE,SUBPARTITIONING_TYPE,PARTITION_COUNT,LOCALITY,DEF_TABLESPACE_NAME from DBA_PART_INDEXES where index_name=‘IND_2‘;
INDEX_NAME TABLE_NAME PARTITION SUBPARTIT PARTITION_COUNT LOCALI DEF_TABLESPACE_NAME
---------- ------------------------------ --------- --------- --------------- ------ ------------------------------
IND_2 RANGE_PART_TAB RANGE NONE 13 LOCAL
三、索引的重建
非分區索引：
SQL> alter index ind_1 rebuild;
全局分區索引：
SQL> alter table range_part_tab move partition p1;
Table altered.
SQL> select TABLE_NAME,PARTITION_NAME,TABLESPACE_NAME from dba_tab_partitions
where TABLE_OWNER=‘SYS‘ and TABLE_NAME=‘RANGE_PART_TAB‘;
TABLE_NAME PARTITION_ TABLESPACE_NAME
------------------------------ ---------- ------------------------------
RANGE_PART_TAB P1 SYSTEM
RANGE_PART_TAB P10 SYSTEM
RANGE_PART_TAB P11 SYSTEM
SQL> select index_name,PARTITION_NAME,HIGH_VALUE,STATUS,TABLESPACE_NAME
from dba_ind_partitions where index_owner=‘SYS‘ and index_name=‘IND_3‘;
INDEX_NAME PARTITION_ HIGH_VALUE STATUS TABLESPACE_NAME
---------- ---------- -------------------- -------- ------------------------------
IND_3 P1 592 USABLE SYSTEM
IND_3 P2 11100 USABLE USERS
IND_3 P3 MAXVALUE USABLE SYSAUX
SQL> alter index ind_3 rebuild partition p1;
Index altered.
全局分區索引，單個分區的move並不會導致全局分區索引的失效
本地分區索引
select index_name,PARTITION_NAME,STATUS,TABLESPACE_NAME
from dba_ind_partitions where index_owner=‘SYS‘ and index_name=‘IND_2‘;
INDEX_NAME PARTITION_ STATUS TABLESPACE_NAME
---------- ---------- -------- ------------------------------
IND_2 P1 USABLE SYSTEM
IND_2 P10 USABLE SYSTEM
IND_2 P11 USABLE SYSTEM
IND_2 P12 USABLE SYSTEM
SQL> alter table range_part_tab move partition p1;
Table altered
SQL> alter index ind_2 rebuild partition p1;
Index altered.
全局（Global) 與本地（Local)索引的區別