Oracle之索引三大特徵

索引**三大特徵:

1.索引的高度較低

索引高度較低的學習：

隨著資料的不斷增多，從底層的block塊開始，如果儲存滿了會向上儲存目錄，但是向上增長緩慢，查詢資料的以後從最上級開始向下查詢，一層乙個io，大量資料主要在吞吐量緩慢。

測試：可以建立表插入資料，然後建立索引，然後用hitns進行全表掃瞄，檢視執行計畫看consistent gets數量。

例子：drop table t1 purge;

drop table t2 purge;

drop table t3 purge;

drop table t4 purge;

drop table t5 purge;

drop table t6 purge;

drop table t7 purge;

create table t1 as select rownum as id ,rownum+1 as id2,rpad('*',1000,'*') as contents from dual connect by level<=1;

create table t2 as select rownum as id ,rownum+1 as id2,rpad('*',1000,'*') as contents from dual connect by level<=10;

create table t3 as select rownum as id ,rownum+1 as id2,rpad('*',1000,'*') as contents from dual connect by level<=100;

create table t4 as select rownum as id ,rownum+1 as id2,rpad('*',1000,'*') as contents from dual connect by level<=1000;

create table t5 as select rownum as id ,rownum+1 as id2,rpad('*',1000,'*') as contents from dual connect by level<=10000;

create table t6 as select rownum as id ,rownum+1 as id2,rpad('*',1000,'*') as contents from dual connect by level<=100000;

create table t7 as select rownum as id ,rownum+1 as id2,rpad('*',1000,'*') as contents from dual connect by level<=1000000;

create index idx_id_t1 on t1(id);

create index idx_id_t2 on t2(id);

create index idx_id_t3 on t3(id);

create index idx_id_t4 on t4(id);

create index idx_id_t5 on t5(id);

create index idx_id_t6 on t6(id);

create index idx_id_t7 on t7(id);

set linesize 1000

set autotrace off

select index_name,

blevel,

leaf_blocks,

num_rows,

distinct_keys,

clustering_factor

from user_ind_statistics

where table_name in( 't1','t2','t3','t4','t5','t6','t7');

2.儲存列值：列值+rowid

3.本身有序

例子：優化count(*),如果列值存在null那麼索引將失效。在列值上加上索引，count(*)會走索引。

補充：sum等也會走索引

例子：select id from t order by id; 檢視執行計畫不會產生sorts(memory)，因為索引本身就是排序的

例子：select max(id) from t;速度異常快，因為不論表的資料如何增加，max min是在固定位置的，所以查詢很快。

例子：union無法優化，索引無法消除union排序，一般來說使用union存在必要性，在資料不會重複時候用union all

例子：索引之排序

drop table t purge;

create table t as select * from dba_objects ;

set autotrace traceonly

--oracle還算智慧型，不會傻到這裡都去排序，做了查詢轉換，忽略了這個排序

select count(*) from t order by object_id;

---以下語句說明排序

set autotrace traceonly

set linesize 1000

drop table t purge;

create table t as select * from dba_objects;

--以下語句沒有索引又有order by ，必然產生排序

select * from t where object_id>2 order by object_id;

---新增索引後，oracle就有可能利用索引本身就有序的特點，利用索引來避免排序，如下：

create index idx_t_object_id on t(object_id);

set autotrace traceonly

select * from t where object_id>2 order by object_id;

--如下情況oracle肯定毫不猶豫的選擇用索引,因為回表取消了 !

select object_id from t where object_id>2 order by object_id;

例子：索引值max/min查詢

--max/min 的索引優化

drop table t purge;

create table t as select * from dba_objects;

update t set object_id=rownum;

alter table t add constraint pk_object_id primary key (object_id);

set autotrace on

set linesize 1000

select max(object_id) from t;

--最小值老師的試驗就無需展現執行計畫結果了，必然和最大值的執行計畫一樣！

select min(object_id) from t;

--如果沒用到索引的情況是如下,請看看執行計畫有何不同，請看看代價和邏輯讀的差異！

select /*+full(t)*/ max(object_id) from t;

---另外，可以做如下試驗觀察在有索引的情況下，隨這記錄數增加，效能差異是否明顯？

set autotrace off

drop table t_max purge;

create table t_max as select * from dba_objects;

insert into t_max select * from t_max;

select count(*) from t_max;

create index idx_t_max_obj on t_max(object_id);

set autotrace on

select max(object_id) from t_max;

補充知識：

index full scan (min/max)

table access by index rowid（回表）

Oracle之索引三大特徵

C 三大特徵之繼承

oop三大特徵之繼承

物件導向三大特徵之封裝

Oracle之索引三大特徵

C 三大特徵之繼承

oop三大特徵之繼承

物件導向三大特徵之封裝

相關推薦