3, oracle table space and index operations


3.1, create a table space and user authorization:

1, create a table space:

CREATE TABLESPACE <tablespace name> LOGGING DATAFILE '<storage path>' SIZE 50M AUTOEXTEND ON NEXT 50M MAXSIZE 31768M EXTENT MANAGEMENT LOCAL;

#windows storage path: D: \ app \ Administrator \ oradata \ orcl \ lc_data.dbf

#linux storage path: /application/oracle/oradata/orcl/lc_data.dbf

2, create a user and specify the table space:

CREATE USER <username> IDENTIFIED BY <password> DEFAULT TABLESPACE <tablespace name>;

# A user has only one table space, and table space can have multiple users;

3, grant permissions to the user:

grant connect, resource, dba to <username>;

3.2, delete the table space:

1, delete users:

drop user <user name> cascade;

2, delete the table space:

drop tablespace <表空间名> including contents and datafiles cascade constraints;

#including contents: Delete the contents of the table space before you drop a table space table space in the content, without adding this parameter, the table space can not be deleted;

#including datafiles: data files to delete the table space;

#cascade constraints: foreign key to delete reference to tables in the tablespace;

3.3, table space query:

1, the query all the table space and the corresponding path:

select tablespace_name,file_name from dba_data_files;

2, query the status of all table space information;

select tablespace_name,status from dba_tablespaces;

3, increase table space size:

ALTER TABLESPACE <tablespace name> ADD DATAFILE '<storage path>' SIZE 50M AUTOEXTEND ON NEXT 50M MAXSIZE UNLIMITED;

4, view all users and the corresponding table space:

select username,default_tablespace from dba_users;

5, view the current user's default table space:

select username,default_tablespace from user_users;

6, the link between the table space:

(1) create a link between the table space:

create database link TO_ <tablespace name> _LINK connect to <username> identified by <password> using '<database instance name>';

(2) the link between the look-up table space:

select * from <current tablespace name> @TO_ <tablespace name> _LINK;

(3) the link between the drop table space:

drop database link TO_ <tablespace name> _LINK;

7, all the amount of table space query:

select tablespace_name,count(*) AS extends,round(sum(bytes)/1024/1024,2) AS MB,sum(blocks) AS blocks from dba_free_space group BY tablespace_name;

8, see the serial number, last_number is the current value:

select * from user_sequences;

9, to modify the table space attributes (offline):

alter tablespace <tablespace name> offline;

10、修改表空间属性(在线):

alter tablespace <表空间名> online;

11、修改表空间属性(只读):

alter tablespace <表空间名> read only;

12、修改表空间属性(读写)

alter tablespace <表空间名> read write;

13、修改session的时间格式:

alter session set nls_date_format='yyyy-mm-dd';

3.4、索引操作:

1、创建单一索引:

create index <索引名> on <表名>(<列名1>);

2、创建组合索引:

create index <索引名> on <表名>(<列名1>,<列名2>);

select * from <表名> where <列名1>='<字符>'

#走索引

select * from <表名> where <列名2>='<字符>'

#不走索引

select * from <表名> where <列名1> like '%<字符>%'

#不走索引

select * from <表名> where <列名1>='<字符>' and <列名2>='<字符2>'

#走索引

select * from <表名> where <列名1>='<字符>' or <列名2>='<字符2>'

#不走索引

drop index <索引名称>;

#删除索引

3、查看索引的方法:

(1)在当前用户中查找表名:

select * from user_tables where table_name like '<表名>%';

(2)查询该表的所有索引:

select * from user_indexes where table_name='<表名>';

(3)查询该表的所有索引列:

select * from user_ind_columns where table_name='<表名>';

(4)查询当前用户所有表的索引和索引类别:

select table_name,index_name,index_type from user_indexes order by index_name;

(5)查看当前用户下指定索引的信息:

select * from user_indexes where index_name=upper('&index_name');

(6)查看当前用户下指定的索引的索引列:

select * from user_ind_columns where index_name=upper('&index_name');

(7)查看当前用户下指定索引的大小:

select sum(bytes)/(1024*1024) as "size(M)" from user_segments where segment_name=upper('&index_name');

(8)补充:

user_indexes #存放着当前用户所有表的索引信息;

user_segments #存放着当前用户所有表的索引大小;

user_ind_columns #存放着当前用户所有表的索引列信息;

4、索引补充:

(1)哪些列适合建索引:

1)经常出现在where子句的列;

2)经常用于表连接的列,在匹配表上进行建索引;

3)该列是高基数数据列,高基数数据列是指有很多不同的值;

4)索引里面不计null值;

5)表很大,查询结果集小;

6)在pk、uk、fk键上建立索引;

7)经常需要排序"order by"和分组"group by"的列;

(2)索引用不了的写法:

1)函数导致索引用不了 where upper(colname)= 'char';

2)可以对函数建索引:

create index <索引名> on <表名>(round(<列名1>));

3)表达式导致索引用不了 where colname*12=1200;

4)索引不是万能的;

(3)索引结构:

1)分析索引结构有效性:

analyze index <索引> validate structure;

一般来讲默认的方式是offline;

当以offline的模式analyze索引时,会对table加一个表级共享锁,对目前table的一些实时DMl操作会产生一定的影响;

而以online模式分析时候,则不会加任何lock,但在index_stats中是看不到任何信息的;

2)查看索引结构:

select NAME,HEIGHT,BLOCKS,BR_BLKS,BR_ROWS,LF_BLKS,LF_ROWS from index_stats;

3)合并索引叶级块碎片:

alter index <索引名> coalesce;

4)重建索引:

alter index <索引名> rebuild;






Guess you like

Origin www.cnblogs.com/LiuChang-blog/p/12315505.html