Pular para o conteúdo principal

Indíces Virtuais

Acredito que todos já precisaram criar algum indice para teste de performance na consulta, porém isso pode demandar tempo para indexar uma tabela grande sem contar espaço em disco, ou você não sabe os efeitos de criar um indice e afetar outras consultas do seu banco,
Dado esse fato, podemos utilizar os indices virtuais, ou indices sem segmentos, abaixo um exemplo de como funciona:


SQL> create table tab_vindex as (select * from dba_objects where object_id is not null);

Table created.

SQL> ALTER TABLE tab_vindex add constraint pk_tab_vindex PRIMARY KEY (object_id);

Table altered.

SQL> EXEC DBMS_STATS.gather_table_stats(USER, 'tab_vindex', cascade=>TRUE);

PL/SQL procedure successfully completed.

SQL> SET AUTOTRACE TRACEONLY EXP

SQL> SELECT * FROM tab_vindex where object_id = 100;

Execution Plan
----------------------------------------------------------
Plan hash value: 2213869000

---------------------------------------------------------------------------------------------
| Id | Operation | Name | Rows | Bytes | Cost (%CPU)| Time |
---------------------------------------------------------------------------------------------
| 0 | SELECT STATEMENT | | 1 | 100 | 2 (0)| 00:00:01 |
| 1 | TABLE ACCESS BY INDEX ROWID| TAB_VINDEX | 1 | 100 | 2 (0)| 00:00:01 |
|* 2 | INDEX UNIQUE SCAN | PK_TAB_VINDEX | 1 | | 1 (0)| 00:00:01 |
---------------------------------------------------------------------------------------------

Predicate Information (identified by operation id):
---------------------------------------------------

2 - access("OBJECT_ID"=100)

SQL> SELECT * FROM tab_vindex where object_name = 'TAB_VINDEX';

Execution Plan
----------------------------------------------------------
Plan hash value: 4059569637

--------------------------------------------------------------------------------
| Id | Operation | Name | Rows | Bytes | Cost (%CPU)| Time |
--------------------------------------------------------------------------------
| 0 | SELECT STATEMENT | | 2 | 200 | 528 (1)| 00:00:02 |
|* 1 | TABLE ACCESS FULL| TAB_VINDEX | 2 | 200 | 528 (1)| 00:00:02 |
--------------------------------------------------------------------------------

Predicate Information (identified by operation id):
---------------------------------------------------

1 - filter("OBJECT_NAME"='TAB_VINDEX')

SQL> CREATE INDEX ix_tab_vindex ON tab_vindex(object_name) NOSEGMENT;

Index created.

SQL> SELECT * FROM tab_vindex where object_name = 'TAB_VINDEX';

Execution Plan
----------------------------------------------------------
Plan hash value: 4059569637

--------------------------------------------------------------------------------
| Id | Operation | Name | Rows | Bytes | Cost (%CPU)| Time |
--------------------------------------------------------------------------------
| 0 | SELECT STATEMENT | | 2 | 200 | 528 (1)| 00:00:02 |
|* 1 | TABLE ACCESS FULL| TAB_VINDEX | 2 | 200 | 528 (1)| 00:00:02 |
--------------------------------------------------------------------------------

Predicate Information (identified by operation id):
---------------------------------------------------

1 - filter("OBJECT_NAME"='TAB_VINDEX')

PARA FUNCIONAR É NECESSÁRIO SETAR O PARAMETRO: _use_nosegment_indexes

SQL> ALTER SESSION SET "_use_nosegment_indexes" = TRUE;

Session altered.

SQL> SELECT * FROM tab_vindex where object_name = 'TAB_VINDEX';

Execution Plan
----------------------------------------------------------
Plan hash value: 1624739785

---------------------------------------------------------------------------------------------
| Id | Operation | Name | Rows | Bytes | Cost (%CPU)| Time |
---------------------------------------------------------------------------------------------
| 0 | SELECT STATEMENT | | 2 | 200 | 2 (0)| 00:00:01 |
| 1 | TABLE ACCESS BY INDEX ROWID| TAB_VINDEX | 2 | 200 | 2 (0)| 00:00:01 |
|* 2 | INDEX RANGE SCAN | IX_TAB_VINDEX | 2 | | 1 (0)| 00:00:01 |
---------------------------------------------------------------------------------------------

Predicate Information (identified by operation id):
---------------------------------------------------

2 - access("OBJECT_NAME"='TAB_VINDEX')

SQL> SET AUTOTRACE OFF

SQL> SELECT index_name FROM user_indexes;

INDEX_NAME
------------------------------
PK_TAB_VINDEX

SQL> SELECT object_name FROM user_objects WHERE object_type = 'INDEX';

OBJECT_NAME
--------------------------------------------------------------------------------------------------------------------------------
PK_TAB_VINDEX
IX_TAB_VINDEX

SQL> spool off

Comentários

  1. O uso de Indíces Virtuais é um recurso extremamente interessante e pouco utilizado, não sei porquê, digo pouco utilizado porque já trabalhei com ótimos dba's e vi poucos, pra nao dizer quase nenhum utilizar ou citar.

    òtimo o seu blog

    ResponderExcluir
  2. Obrigado Hudson... acho uma excelente feature e para análise em sistemas transacionais é muito útil.

    ResponderExcluir

Postar um comentário

Postagens mais visitadas deste blog

Index SKIP SCAN???

Index Skip Scan nada mais é que o "split" logico de um indice composto, ou seja, se sua clausula WHERE não tiver todos os campos do indice composto, ele é dividido em menores sub-indices Exemplo: -------- SQL> create table teste(a number,b number); Table created. SQL> begin 2 for i in 1..1000 3 loop 4 insert into teste values(i, 56); 5 end loop; 6 end; 7 / PL/SQL procedure successfully completed. SQL> create index idx_teste on teste(a,b); Index created. SQL> analyze table teste compute statistics; Table analyzed. SQL> explain plan for select /*+ index_ss(teste idx_teste) */ * from teste where b=56; Explained. SQL> start ?\rdbms\admin\utlxpls.sql -------------------------------------------------------------------------------- | Operation | Name | Rows | Bytes| Cost | Pstart| Pstop | -------------------------------------------------------------------------------- | SELECT STATEMENT | | 1

Webcenter Content - Jython to Checkin Content with Attachments (CHECKIN_NEW_WITH_RENDITIONS)

In the last few days, i had to load a content that, by its requirements, needs to attach files to the content. My first alternative would be Batch Loading, however for the bacth loder file it was not possible because it does not actually support the rendition service. With the next option, i tried to use the IdcCommand, but still had many problems, such as errors below: sample hda file: [oracle @ hostname bin] $ cat idctst.txt @Properties LocalData IdcService = EDIT_RENDITIONS dID = 37020 renditionKeys = addRendition0 addRendition0.name = test543 addRendition0.action = edit addRendition0.file = test.txt addRendition0.file: path = / u01 / oracle / domains / content_domain / ucm / cs / vault / ~ temp / test.txt @end << EOD >> Executing: [oracle @ hostname bin] $ ./IdcCommand -f idctst.txt -u sysadmin -l log.log -c server Error: Java HotSpot (TM) 64-Bit Server VM warning: Using the ParNew young collector with the Serial old collector is deprecated and w

Configurar Tamanho de UNDO e tempo de RETENÇÃO (UNDO_RETENTION)

-- Tamanho Undo Atual SELECT SUM(a.bytes) "UNDO_SIZE" FROM v$datafile a, v$tablespace b, dba_tablespaces c WHERE c.contents = 'UNDO' AND c.status = 'ONLINE' AND b.name = c.tablespace_name AND a.ts# = b.ts#; -- Blocos de UNDO por Seg. SELECT MAX(undoblks/((end_time-begin_time)*3600*24)) "UNDO_BLOCK_PER_SEC" FROM v$undostat; --Tamanho do Bloco SELECT TO_NUMBER(value) "DB_BLOCK_SIZE [KByte]" FROM v$parameter WHERE name = 'db_block_size'; -- Select calcula UNDO RETENTION SELECT d.undo_size/(1024*1024) "ACTUAL UNDO SIZE [MByte]", SUBSTR(e.value,1,25) "UNDO RETENTION [Sec]", ROUND((d.undo_size / (to_number(f.value) * g.undo_block_per_sec))) "OPTIMAL UNDO RETENTION [Sec]" FROM ( SELECT SUM(a.bytes) undo_size FROM v$datafile a, v$tablespace b, dba_tablespaces c WHERE c.contents = 'UNDO'