{"id":191,"date":"2015-12-20T22:47:22","date_gmt":"2015-12-21T00:47:22","guid":{"rendered":"http:\/\/dbtimewizard.com.br\/blog\/?p=191"},"modified":"2019-10-10T23:10:23","modified_gmt":"2019-10-11T02:10:23","slug":"execution-plan-como-interpretar-as-secoes-nao-estatisticas-do-plano-de-execucao","status":"publish","type":"post","link":"http:\/\/dbtimewizard.com.br\/blog\/execution-plan-como-interpretar-as-secoes-nao-estatisticas-do-plano-de-execucao\/","title":{"rendered":"Execution Plan: Como interpretar as se\u00e7\u00f5es n\u00e3o estat\u00edsticas do plano de execu\u00e7\u00e3o"},"content":{"rendered":"

Nas publica\u00e7\u00f5es anteriores foram abordados os t\u00f3picos: como determinar a sequ\u00eancia de execu\u00e7\u00e3o do plano<\/a><\/strong>, como interpretar a estat\u00edsticas estimadas<\/a> <\/strong>e as estat\u00edsticas geradas durante a execu\u00e7\u00e3o do plano<\/a><\/strong>, agora vamos abordar algumas se\u00e7\u00f5es complementares que fornecem informa\u00e7\u00f5es importantes para entender por que o plano de execu\u00e7\u00e3o foi criado com determinadas opera\u00e7\u00f5es, por que essas opera\u00e7\u00f5es utilizam determinados recursos e por que estas opera\u00e7\u00f5es foram ordenadas numa sequ\u00eancia especifica.<\/span><\/p>\n

Query Block Name \/ Object Alias Section<\/span><\/h2>\n

Uma instru\u00e7\u00e3o SQL pode ser composta de v\u00e1rias sub-consultas\u00a0e as vezes \u00e9 dif\u00edcil fazer uma correla\u00e7\u00e3o entre a instru\u00e7\u00e3o SQL e o plano de execu\u00e7\u00e3o, para facilitar esta leitura podemos utilizar esta se\u00e7\u00e3o e atrav\u00e9s do “Object Alias” identificar no plano cada umas das sub-consultas.<\/span><\/p>\n

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
<\/div><\/td>
SELECT *<\/span> FROM V$VERSION<\/span> where rownum <<\/span> 2<\/span>;<\/span>
\n
\nBANNER
\n--------------------------------------------------------------------------------<\/span>
\nOracle Database 11g Enterprise Edition Release 11.<\/span>2.<\/span>0.<\/span>4.<\/span>0 -<\/span> 64bit Production
\n
\n
\nSELECT outer.*<\/span>
\n  FROM employees outer
\n WHERE outer.<\/span>salary ><\/span> (<\/span>SELECT avg(<\/span>inner.<\/span>salary)<\/span>
\n                         FROM employees inner
\n                        WHERE inner.<\/span>department_id =<\/span> outer.<\/span>department_id)<\/span>;<\/span>
\n
\n
\nSELECT *<\/span> FROM TABLE(<\/span>dbms_xplan.<\/span>display_cursor(<\/span>'b2dtq04ypz3kw'<\/span>,<\/span>0<\/span>,<\/span>'+alias'<\/span>)<\/span>)<\/span>;<\/span>
\n
\n----------------------------------------------------------------------------------<\/span>
\n|<\/span> Id  |<\/span> Operation            |<\/span> Name      |<\/span> Rows  |<\/span> Bytes |<\/span> Cost (<\/span>%<\/span>CPU)<\/span>|<\/span> Time<\/span><\/a>     |<\/span>
\n----------------------------------------------------------------------------------<\/span>
\n|<\/span>   0<\/span> |<\/span> SELECT STATEMENT     |<\/span>           |<\/span>       |<\/span>       |<\/span>     7<\/span> (<\/span>100<\/span>)<\/span>|<\/span>          |<\/span>
\n|*<\/span>  1<\/span> |<\/span>  HASH<\/span><\/a> JOIN<\/span><\/a>           |<\/span>           |<\/span>    17<\/span> |<\/span>  1615<\/span> |<\/span>     7<\/span>  (<\/span>15<\/span>)<\/span>|<\/span> 00<\/span>:<\/span>00<\/span>:<\/span>01<\/span> |<\/span>
\n|<\/span>   2<\/span> |<\/span>   VIEW               |<\/span> VW_SQ_1   |<\/span>    11<\/span> |<\/span>   286<\/span> |<\/span>     4<\/span>  (<\/span>25<\/span>)<\/span>|<\/span> 00<\/span>:<\/span>00<\/span>:<\/span>01<\/span> |<\/span>
\n|<\/span>   3<\/span> |<\/span>    HASH<\/span><\/a> GROUP BY     |<\/span>           |<\/span>    11<\/span> |<\/span>    77<\/span> |<\/span>     4<\/span>  (<\/span>25<\/span>)<\/span>|<\/span> 00<\/span>:<\/span>00<\/span>:<\/span>01<\/span> |<\/span>
\n|<\/span>   4<\/span> |<\/span>     TABLE ACCESS FULL|<\/span> EMPLOYEES |<\/span>   107<\/span> |<\/span>   749<\/span> |<\/span>     3<\/span>   (<\/span>0<\/span>)<\/span>|<\/span> 00<\/span>:<\/span>00<\/span>:<\/span>01<\/span> |<\/span>
\n|<\/span>   5<\/span> |<\/span>   TABLE ACCESS FULL  |<\/span> EMPLOYEES |<\/span>   107<\/span> |<\/span>  7383<\/span> |<\/span>     3<\/span>   (<\/span>0<\/span>)<\/span>|<\/span> 00<\/span>:<\/span>00<\/span>:<\/span>01<\/span> |<\/span>
\n----------------------------------------------------------------------------------<\/span>
\n
\nQuery Block Name \/<\/span> Object Alias (<\/span>identified by operation id)<\/span>:<\/span>
\n-------------------------------------------------------------<\/span>
\n
\n   1<\/span> -<\/span> SEL$C772B8D1<\/span>
\n   2<\/span> -<\/span> SEL$683B0107 \/<\/span> VW_SQ_1@<\/span>SEL$7511BFD2
\n   3<\/span> -<\/span> SEL$683B0107
\n   4<\/span> -<\/span> SEL$683B0107 \/<\/span> INNER@<\/span>SEL$2<\/span>
\n   5<\/span> -<\/span> SEL$C772B8D1<\/span> \/<\/span> OUTER@<\/span>SEL$1<\/span><\/div><\/td><\/tr><\/tbody><\/table><\/div>\n

Na instru\u00e7\u00e3o SQL acima temos uma consulta principal e na clausula WHERE temos uma sub-consulta, se observarmos o plano de execu\u00e7\u00e3o fica dif\u00edcil determinar quais linhas se referem a sub-consulta, mas observando o “Object Alias” da linha 4 na se\u00e7\u00e3o “Query Block Name” SEL$683B0107 podemos ver que foi utilizado o “Alias” da tabela da sub-consulta “INNER” para compor o nome do “Object Alias” INNER@SEL$2, desta forma podemos dizer que as linha 2,3,4 que possuem o mesmo QB_NAME SEL$683B0107 s\u00e3o opera\u00e7\u00f5es referentes a sub-consulta e as linhas 1 e 5 cujo QB_NAME \u00e9 SEL$C772B8D1 s\u00e3o opera\u00e7\u00f5es referentes a consulta principal.<\/span><\/p>\n

Quando temos v\u00e1rias sub-consultas referenciando a mesma tabela e n\u00e3o foi utilizado “Alias” para identificar as tabelas fica dif\u00edcil identificar as “Query Block”, neste caso podemos utilizar o “Hint QB_NAME” que vai facilitar muito essa identifica\u00e7\u00e3o. A seguir temo um exemplo de utiliza\u00e7\u00e3o desse “Hint”.<\/span><\/p>\n

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
<\/div><\/td>
SELECT (<\/span>SELECT \/*+ QB_NAME(salesman) *\/<\/span> COUNT<\/span><\/a>(<\/span>*<\/span>)<\/span> FROM employees WHERE job_id like '%REP'<\/span>)<\/span> AS<\/span> salesman_count,<\/span>
\n       (<\/span>SELECT \/*+ QB_NAME(manager) *\/<\/span>  COUNT<\/span><\/a>(<\/span>*<\/span>)<\/span> FROM employees WHERE job_id like '%MAN'<\/span>)<\/span> AS<\/span> manager_count
\nFROM   dual;<\/span>
\n
\nSELECT *<\/span> FROM TABLE(<\/span>dbms_xplan.<\/span>display_cursor(<\/span>'gs5d0xtyta4zg'<\/span>,<\/span>0<\/span>,<\/span>'+alias'<\/span>)<\/span>)<\/span>
\n
\n--------------------------------------------------------------------------------<\/span>
\n|<\/span> Id  |<\/span> Operation         |<\/span> Name       |<\/span> Rows  |<\/span> Bytes |<\/span> Cost (<\/span>%<\/span>CPU)<\/span>|<\/span> Time<\/span><\/a>     |<\/span>
\n--------------------------------------------------------------------------------<\/span>
\n|<\/span>   0<\/span> |<\/span> SELECT STATEMENT  |<\/span>            |<\/span>       |<\/span>       |<\/span>     2<\/span> (<\/span>100<\/span>)<\/span>|<\/span>          |<\/span>
\n|<\/span>   1<\/span> |<\/span>  SORT<\/span><\/a> AGGREGATE<\/span><\/a>   |<\/span>            |<\/span>     1<\/span> |<\/span>     9<\/span> |<\/span>            |<\/span>          |<\/span>
\n|*<\/span>  2<\/span> |<\/span>   INDEX RANGE<\/span><\/a> SCAN|<\/span> EMP_JOB_IX |<\/span>     5<\/span> |<\/span>    45<\/span> |<\/span>     1<\/span>   (<\/span>0<\/span>)<\/span>|<\/span> 00<\/span>:<\/span>00<\/span>:<\/span>01<\/span> |<\/span>
\n|<\/span>   3<\/span> |<\/span>  SORT<\/span><\/a> AGGREGATE<\/span><\/a>   |<\/span>            |<\/span>     1<\/span> |<\/span>     9<\/span> |<\/span>            |<\/span>          |<\/span>
\n|*<\/span>  4<\/span> |<\/span>   INDEX RANGE<\/span><\/a> SCAN|<\/span> EMP_JOB_IX |<\/span>     5<\/span> |<\/span>    45<\/span> |<\/span>     1<\/span>   (<\/span>0<\/span>)<\/span>|<\/span> 00<\/span>:<\/span>00<\/span>:<\/span>01<\/span> |<\/span>
\n|<\/span>   5<\/span> |<\/span>  FAST DUAL        |<\/span>            |<\/span>     1<\/span> |<\/span>       |<\/span>     2<\/span>   (<\/span>0<\/span>)<\/span>|<\/span> 00<\/span>:<\/span>00<\/span>:<\/span>01<\/span> |<\/span>
\n--------------------------------------------------------------------------------<\/span>
\n
\nQuery Block Name \/<\/span> Object Alias (<\/span>identified by operation id)<\/span>:<\/span>
\n-------------------------------------------------------------<\/span>
\n
\n   1<\/span> -<\/span> SALESMAN
\n   2<\/span> -<\/span> SALESMAN \/<\/span> EMPLOYEES@<\/span>SALESMAN
\n   3<\/span> -<\/span> MANAGER
\n   4<\/span> -<\/span> MANAGER  \/<\/span> EMPLOYEES@<\/span>MANAGER
\n   5<\/span> -<\/span> SEL$1<\/span>    \/<\/span> DUAL@<\/span>SEL$1<\/span><\/div><\/td><\/tr><\/tbody><\/table><\/div>\n

Outline Data Section<\/span><\/h2>\n

A se\u00e7\u00e3o “Outline Data” \u00e9 uma lista de “hints” que pode ser utilizada para replicar um plano de execu\u00e7\u00e3o mesmo que as estat\u00edsticas das tabelas sejam alteradas, esta lista pode ser copiada para a instru\u00e7\u00e3o SQL sem altera\u00e7\u00f5es. A seguir um exemplo de como podemos listar esta se\u00e7\u00e3o utilizando o pacote DBMS_XPLAN.<\/span><\/p>\n

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
<\/div><\/td>
SELECT outer.*<\/span>
\n  FROM employees outer
\n WHERE outer.<\/span>salary ><\/span> (<\/span>SELECT avg(<\/span>inner.<\/span>salary)<\/span>
\n                         FROM employees inner
\n                        WHERE inner.<\/span>DEPARTMENT_ID =<\/span> outer.<\/span>DEPARTMENT_ID)<\/span>;<\/span>
\n
\nSELECT *<\/span> FROM TABLE(<\/span>dbms_xplan.<\/span>display_cursor(<\/span>'b2dtq04ypz3kw'<\/span>,<\/span>0<\/span>,<\/span>'+outline'<\/span>)<\/span>)<\/span>
\n
\n
\n----------------------------------------------------------------------------------<\/span>
\n|<\/span> Id  |<\/span> Operation            |<\/span> Name      |<\/span> Rows  |<\/span> Bytes |<\/span> Cost (<\/span>%<\/span>CPU)<\/span>|<\/span> Time<\/span><\/a>     |<\/span>
\n----------------------------------------------------------------------------------<\/span>
\n|<\/span>   0<\/span> |<\/span> SELECT STATEMENT     |<\/span>           |<\/span>       |<\/span>       |<\/span>     7<\/span> (<\/span>100<\/span>)<\/span>|<\/span>          |<\/span>
\n|*<\/span>  1<\/span> |<\/span>  HASH<\/span><\/a> JOIN<\/span><\/a>           |<\/span>           |<\/span>    17<\/span> |<\/span>  1615<\/span> |<\/span>     7<\/span>  (<\/span>15<\/span>)<\/span>|<\/span> 00<\/span>:<\/span>00<\/span>:<\/span>01<\/span> |<\/span>
\n|<\/span>   2<\/span> |<\/span>   VIEW               |<\/span> VW_SQ_1   |<\/span>    11<\/span> |<\/span>   286<\/span> |<\/span>     4<\/span>  (<\/span>25<\/span>)<\/span>|<\/span> 00<\/span>:<\/span>00<\/span>:<\/span>01<\/span> |<\/span>
\n|<\/span>   3<\/span> |<\/span>    HASH<\/span><\/a> GROUP BY     |<\/span>           |<\/span>    11<\/span> |<\/span>    77<\/span> |<\/span>     4<\/span>  (<\/span>25<\/span>)<\/span>|<\/span> 00<\/span>:<\/span>00<\/span>:<\/span>01<\/span> |<\/span>
\n|<\/span>   4<\/span> |<\/span>     TABLE ACCESS FULL|<\/span> EMPLOYEES |<\/span>   107<\/span> |<\/span>   749<\/span> |<\/span>     3<\/span>   (<\/span>0<\/span>)<\/span>|<\/span> 00<\/span>:<\/span>00<\/span>:<\/span>01<\/span> |<\/span>
\n|<\/span>   5<\/span> |<\/span>   TABLE ACCESS FULL  |<\/span> EMPLOYEES |<\/span>   107<\/span> |<\/span>  7383<\/span> |<\/span>     3<\/span>   (<\/span>0<\/span>)<\/span>|<\/span> 00<\/span>:<\/span>00<\/span>:<\/span>01<\/span> |<\/span>
\n----------------------------------------------------------------------------------<\/span>
\n
\nOutline Data
\n-------------<\/span>
\n
\n  \/*+
\n      BEGIN_OUTLINE_DATA
\n      IGNORE_OPTIM_EMBEDDED_HINTS
\n      OPTIMIZER_FEATURES_ENABLE('11.2.0.4')
\n      DB_VERSION('11.2.0.4')
\n      ALL_ROWS
\n      OUTLINE_LEAF(@"SEL$683B0107")
\n      OUTLINE_LEAF(@"SEL$C772B8D1")
\n      UNNEST(@"SEL$2")
\n      OUTLINE(@"SEL$2")
\n      OUTLINE(@"SEL$7511BFD2")
\n      OUTLINE(@"SEL$1")
\n      NO_ACCESS(@"SEL$C772B8D1" "VW_SQ_1"@"SEL$7511BFD2")
\n      FULL(@"SEL$C772B8D1" "OUTER"@"SEL$1")
\n      LEADING(@"SEL$C772B8D1" "VW_SQ_1"@"SEL$7511BFD2" "OUTER"@"SEL$1")
\n      USE_HASH(@"SEL$C772B8D1" "OUTER"@"SEL$1")
\n      FULL(@"SEL$683B0107" "INNER"@"SEL$2")
\n      USE_HASH_AGGREGATION(@"SEL$683B0107")
\n      END_OUTLINE_DATA
\n  *\/<\/span><\/div><\/td><\/tr><\/tbody><\/table><\/div>\n

Remote SQL Information Section<\/span><\/h2>\n

A se\u00e7\u00e3o “Remote SQL Information” fornece informa\u00e7\u00f5es complementares a repeito da opera\u00e7\u00e3o “REMOTE” do plano de execu\u00e7\u00e3o sempre que uma instru\u00e7\u00e3o SQL acessar dados de um objeto em outro banco de dados. Nessa se\u00e7\u00e3o vamos encontrar a instru\u00e7\u00e3o SQL que ser\u00e1 submetida na base remota e uma identifica\u00e7\u00e3o do nome do DBLINK utilizado para acessar a base remota.<\/span><\/p>\n

 <\/p>\n

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
<\/div><\/td>
SELECT first_name,<\/span> hire_date,<\/span> department_name
\n  FROM employees e,<\/span>
\n       (<\/span>select department_id,<\/span> department_name from departments@<\/span>APPL2)<\/span> d
\n WHERE e.<\/span>department_id =<\/span> d.<\/span>department_id
\n   AND department_name =<\/span> 'IT'<\/span>;<\/span>
\n
\nSELECT *<\/span> FROM TABLE(<\/span>dbms_xplan.<\/span>display_cursor(<\/span>'08qz6fgcc1cgp'<\/span>,<\/span>0<\/span>,<\/span>'TYPICAL'<\/span>)<\/span>)<\/span>
\n
\n------------------------------------------------------------------------------------------------------------------<\/span>
\n|<\/span> Id  |<\/span> Operation                    |<\/span> Name              |<\/span> Rows  |<\/span> Bytes |<\/span> Cost (<\/span>%<\/span>CPU)<\/span>|<\/span> Time<\/span><\/a>     |<\/span> Inst   |<\/span>IN-<\/span>OUT|<\/span>
\n------------------------------------------------------------------------------------------------------------------<\/span>
\n|<\/span>   0<\/span> |<\/span> SELECT STATEMENT             |<\/span>                   |<\/span>       |<\/span>       |<\/span>     3<\/span> (<\/span>100<\/span>)<\/span>|<\/span>          |<\/span>        |<\/span>      |<\/span>
\n|<\/span>   1<\/span> |<\/span>  NESTED LOOPS                |<\/span>                   |<\/span>    10<\/span> |<\/span>   390<\/span> |<\/span>     3<\/span>   (<\/span>0<\/span>)<\/span>|<\/span> 00<\/span>:<\/span>00<\/span>:<\/span>01<\/span> |<\/span>        |<\/span>      |<\/span>
\n|<\/span>   2<\/span> |<\/span>   NESTED LOOPS               |<\/span>                   |<\/span>    10<\/span> |<\/span>   390<\/span> |<\/span>     3<\/span>   (<\/span>0<\/span>)<\/span>|<\/span> 00<\/span>:<\/span>00<\/span>:<\/span>01<\/span> |<\/span>        |<\/span>      |<\/span>
\n|<\/span>   3<\/span> |<\/span>    REMOTE                    |<\/span> DEPARTMENTS       |<\/span>     1<\/span> |<\/span>    21<\/span> |<\/span>     2<\/span>   (<\/span>0<\/span>)<\/span>|<\/span> 00<\/span>:<\/span>00<\/span>:<\/span>01<\/span> |<\/span>  APPL2 |<\/span> R-><\/span>S<\/span> |<\/span>
\n|*<\/span>  4<\/span> |<\/span>    INDEX RANGE<\/span><\/a> SCAN          |<\/span> EMP_DEPARTMENT_IX |<\/span>    10<\/span> |<\/span>       |<\/span>     0<\/span>   (<\/span>0<\/span>)<\/span>|<\/span>          |<\/span>        |<\/span>      |<\/span>
\n|<\/span>   5<\/span> |<\/span>   TABLE ACCESS BY INDEX ROWID|<\/span> EMPLOYEES         |<\/span>    10<\/span> |<\/span>   180<\/span> |<\/span>     1<\/span>   (<\/span>0<\/span>)<\/span>|<\/span> 00<\/span>:<\/span>00<\/span>:<\/span>01<\/span> |<\/span>        |<\/span>      |<\/span>
\n------------------------------------------------------------------------------------------------------------------<\/span>
\n
\nPredicate Information (<\/span>identified by operation id)<\/span>:<\/span>
\n---------------------------------------------------<\/span>
\n
\n   4<\/span> -<\/span> access(<\/span>"E"<\/span>.<\/span>"DEPARTMENT_ID"<\/span>=<\/span>"DEPARTMENT_ID"<\/span>)<\/span>
\n
\nRemote SQL Information (<\/span>identified by operation id)<\/span>:<\/span>
\n----------------------------------------------------<\/span>
\n
\n   3<\/span> -<\/span> SELECT "DEPARTMENT_ID"<\/span>,<\/span>"DEPARTMENT_NAME"<\/span> FROM "DEPARTMENTS"<\/span> "DEPARTMENTS"<\/span> WHERE
\n       "DEPARTMENT_NAME"<\/span>=:<\/span>1<\/span> (<\/span>accessing 'APPL2'<\/span> )<\/span><\/div><\/td><\/tr><\/tbody><\/table><\/div>\n

 <\/p>\n

Predicate Information Section<\/span><\/h2>\n

A se\u00e7\u00e3o “Predicate Information” fornece informa\u00e7\u00f5es valiosas de como os dados s\u00e3o recuperados do banco de dados, normalmente vamos ver dois tipos de informa\u00e7\u00e3o nessa se\u00e7\u00e3o:<\/span><\/p>\n

ACCESS<\/strong> – significa que vamos utilizar algum \u00edndice para selecionar os registros espec\u00edficos que necessitamos.<\/span><\/p>\n

FILTER<\/strong> – significa que estamos selecionando mais registros do que necessitamos e precisamos excluir registros de acordo com algum crit\u00e9rio informado na clausula WHERE.<\/span><\/p>\n

Vamos ver a seguir alguns exemplos na pr\u00e1tica:<\/span><\/p>\n

 <\/p>\n

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
<\/div><\/td>
select *<\/span> from DEPARTMENTS where DEPARTMENT_NAME =<\/span> 'Administration'<\/span>;<\/span>
\n
\n 
\nSELECT *<\/span> FROM TABLE(<\/span>dbms_xplan.<\/span>display_cursor(<\/span>'277tz6xdc6v80'<\/span>,<\/span>0<\/span>,<\/span>'TYPICAL'<\/span>)<\/span>)<\/span>
\n
\n
\n---------------------------------------------------------------------------------<\/span>
\n|<\/span> Id  |<\/span> Operation         |<\/span> Name        |<\/span> Rows  |<\/span> Bytes |<\/span> Cost (<\/span>%<\/span>CPU)<\/span>|<\/span> Time<\/span><\/a>     |<\/span>
\n---------------------------------------------------------------------------------<\/span>
\n|<\/span>   0<\/span> |<\/span> SELECT STATEMENT  |<\/span>             |<\/span>       |<\/span>       |<\/span>     3<\/span> (<\/span>100<\/span>)<\/span>|<\/span>          |<\/span>
\n|*<\/span>  1<\/span> |<\/span>  TABLE ACCESS FULL|<\/span> DEPARTMENTS |<\/span>     1<\/span> |<\/span>    21<\/span> |<\/span>     3<\/span>   (<\/span>0<\/span>)<\/span>|<\/span> 00<\/span>:<\/span>00<\/span>:<\/span>01<\/span> |<\/span>
\n---------------------------------------------------------------------------------<\/span>
\n
\nPredicate Information (<\/span>identified by operation id)<\/span>:<\/span>
\n---------------------------------------------------<\/span>
\n
\n   1<\/span> -<\/span> filter(<\/span>"DEPARTMENT_NAME"<\/span>=:<\/span>SYS_B_0)<\/span><\/div><\/td><\/tr><\/tbody><\/table><\/div>\n

 <\/p>\n

Neste exemplo na se\u00e7\u00e3o “Predicate Information” temos um “FILTER” na linha 1 do plano de execu\u00e7\u00e3o, ou seja a execu\u00e7\u00e3o dessa instru\u00e7\u00e3o SQL vai realizar uma opera\u00e7\u00e3o “TABLE ACCESS FULL” na tabela DEPARTMENTS lendo todos os registros e excluindo aqueles cujo DEPARTMENT_NAME n\u00e3o seja igual a ‘Administration’.<\/span><\/p>\n

 <\/p>\n

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
<\/div><\/td>
select *<\/span> from DEPARTMENTS where DEPARTMENT_ID =<\/span> 10<\/span>;<\/span>
\n
\nSELECT *<\/span> FROM TABLE(<\/span>dbms_xplan.<\/span>display_cursor(<\/span>'9a2fjq5g05s7p'<\/span>,<\/span>0<\/span>,<\/span>'TYPICAL'<\/span>)<\/span>)<\/span>
\n
\n
\n-------------------------------------------------------------------------------------------<\/span>
\n|<\/span> Id  |<\/span> Operation                   |<\/span> Name        |<\/span> Rows  |<\/span> Bytes |<\/span> Cost (<\/span>%<\/span>CPU)<\/span>|<\/span> Time<\/span><\/a>     |<\/span>
\n-------------------------------------------------------------------------------------------<\/span>
\n|<\/span>   0<\/span> |<\/span> SELECT STATEMENT            |<\/span>             |<\/span>       |<\/span>       |<\/span>     1<\/span> (<\/span>100<\/span>)<\/span>|<\/span>          |<\/span>
\n|<\/span>   1<\/span> |<\/span>  TABLE ACCESS BY INDEX ROWID|<\/span> DEPARTMENTS |<\/span>     1<\/span> |<\/span>    21<\/span> |<\/span>     1<\/span>   (<\/span>0<\/span>)<\/span>|<\/span> 00<\/span>:<\/span>00<\/span>:<\/span>01<\/span> |<\/span>
\n|*<\/span>  2<\/span> |<\/span>   INDEX UNIQUE SCAN         |<\/span> DEPT_ID_PK  |<\/span>     1<\/span> |<\/span>       |<\/span>     0<\/span>   (<\/span>0<\/span>)<\/span>|<\/span>          |<\/span>
\n-------------------------------------------------------------------------------------------<\/span>
\n
\nPredicate Information (<\/span>identified by operation id)<\/span>:<\/span>
\n---------------------------------------------------<\/span>
\n
\n   2<\/span> -<\/span> access(<\/span>"DEPARTMENT_ID"<\/span>=:<\/span>SYS_B_0)<\/span><\/div><\/td><\/tr><\/tbody><\/table><\/div>\n

 <\/p>\n

Neste exemplo na se\u00e7\u00e3o “Predicate Information” temos um “ACCESS” na linha 2 do plano de execu\u00e7\u00e3o, ou seja a execu\u00e7\u00e3o dessa instru\u00e7\u00e3o SQL vai buscar no \u00edndice DEPT_ID_PK todos os registros cujo DEPARTMENT_ID \u00e9 igual 10 e vai passar o ROWID desses registros para opera\u00e7\u00e3o pai na linha 1 do plano. Nesta simula\u00e7\u00e3o\u00a0o \u00edndice \u00e9 uma Primary_Key e nesse caso s\u00f3 podemos ter 1 registro para cada valor.<\/span><\/p>\n

 <\/p>\n

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
<\/div><\/td>
select *<\/span> from DEPARTMENTS where DEPARTMENT_ID ><\/span> 120<\/span> and DEPARTMENT_NAME like '%Control%'<\/span>;<\/span>
\n
\n
\nSELECT *<\/span> FROM TABLE(<\/span>dbms_xplan.<\/span>display_cursor(<\/span>'d78t1gwv11wx6'<\/span>,<\/span>0<\/span>,<\/span>'TYPICAL'<\/span>)<\/span>)<\/span>
\n
\n
\n-------------------------------------------------------------------------------------------<\/span>
\n|<\/span> Id  |<\/span> Operation                   |<\/span> Name        |<\/span> Rows  |<\/span> Bytes |<\/span> Cost (<\/span>%<\/span>CPU)<\/span>|<\/span> Time<\/span><\/a>     |<\/span>
\n-------------------------------------------------------------------------------------------<\/span>
\n|<\/span>   0<\/span> |<\/span> SELECT STATEMENT            |<\/span>             |<\/span>       |<\/span>       |<\/span>     2<\/span> (<\/span>100<\/span>)<\/span>|<\/span>          |<\/span>
\n|*<\/span>  1<\/span> |<\/span>  TABLE ACCESS BY INDEX ROWID|<\/span> DEPARTMENTS |<\/span>     1<\/span> |<\/span>    21<\/span> |<\/span>     2<\/span>   (<\/span>0<\/span>)<\/span>|<\/span> 00<\/span>:<\/span>00<\/span>:<\/span>01<\/span> |<\/span>
\n|*<\/span>  2<\/span> |<\/span>   INDEX RANGE<\/span><\/a> SCAN          |<\/span> DEPT_ID_PK  |<\/span>    16<\/span> |<\/span>       |<\/span>     1<\/span>   (<\/span>0<\/span>)<\/span>|<\/span> 00<\/span>:<\/span>00<\/span>:<\/span>01<\/span> |<\/span>
\n-------------------------------------------------------------------------------------------<\/span>
\n
\nPredicate Information (<\/span>identified by operation id)<\/span>:<\/span>
\n---------------------------------------------------<\/span>
\n
\n   1<\/span> -<\/span> filter(<\/span>"DEPARTMENT_NAME"<\/span> LIKE :<\/span>SYS_B_1)<\/span>
\n   2<\/span> -<\/span> access(<\/span>"DEPARTMENT_ID"<\/span>>:<\/span>SYS_B_0)<\/span><\/div><\/td><\/tr><\/tbody><\/table><\/div>\n

 <\/p>\n

Neste exemplo na se\u00e7\u00e3o “Predicate Information” temos um “ACCESS” na linha 2 e um “FILTER” na linha 1 do plano de execu\u00e7\u00e3o, levando em considera\u00e7\u00e3o que a opera\u00e7\u00e3o filha sempre ser\u00e1 executada primeiro, a execu\u00e7\u00e3o dessa instru\u00e7\u00e3o SQL vai buscar no \u00edndice DEPT_ID_PK todos os registros cujo DEPARTMENT_ID seja maior que 120 e vai passar os ROWID desses registros para a opera\u00e7\u00e3o pai na linha 1 que vai excluir todos os registros que n\u00e3o tenham a palavra “Control” no campo DEPARTMENT_NAME.<\/p>\n

Column Projection Information Section<\/span><\/h2>\n

A se\u00e7\u00e3o “Column Projection Information” mostra quais s\u00e3o as colunas das tabelas que est\u00e3o sendo trabalhadas em cada linha de opera\u00e7\u00e3o do plano de execu\u00e7\u00e3o, seus respectivos tipos e tamanhos. A seguir um exemplo de como essa se\u00e7\u00e3o \u00e9 exibida.<\/span><\/p>\n

 <\/p>\n

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
<\/div><\/td>
SELECT first_name,<\/span> hire_date,<\/span> department_name
\n  FROM employees e,<\/span>
\n       departments d
\n WHERE e.<\/span>department_id =<\/span> d.<\/span>department_id
\n   AND department_name =<\/span> 'IT'<\/span>;<\/span>
\n
\nSELECT *<\/span> FROM TABLE(<\/span>dbms_xplan.<\/span>display_cursor(<\/span>'apj7zuzx3v4w7'<\/span>,<\/span>0<\/span>,<\/span>'TYPICAL +PROJECTION -PREDICATE'<\/span>)<\/span>)<\/span>
\n
\n--------------------------------------------------------------------------------------------------<\/span>
\n|<\/span> Id  |<\/span> Operation                    |<\/span> Name              |<\/span> Rows  |<\/span> Bytes |<\/span> Cost (<\/span>%<\/span>CPU)<\/span>|<\/span> Time<\/span><\/a>     |<\/span>
\n--------------------------------------------------------------------------------------------------<\/span>
\n|<\/span>   0<\/span> |<\/span> SELECT STATEMENT             |<\/span>                   |<\/span>       |<\/span>       |<\/span>     4<\/span> (<\/span>100<\/span>)<\/span>|<\/span>          |<\/span>
\n|<\/span>   1<\/span> |<\/span>  NESTED LOOPS                |<\/span>                   |<\/span>    10<\/span> |<\/span>   340<\/span> |<\/span>     4<\/span>   (<\/span>0<\/span>)<\/span>|<\/span> 00<\/span>:<\/span>00<\/span>:<\/span>01<\/span> |<\/span>
\n|<\/span>   2<\/span> |<\/span>   NESTED LOOPS               |<\/span>                   |<\/span>    10<\/span> |<\/span>   340<\/span> |<\/span>     4<\/span>   (<\/span>0<\/span>)<\/span>|<\/span> 00<\/span>:<\/span>00<\/span>:<\/span>01<\/span> |<\/span>
\n|<\/span>   3<\/span> |<\/span>    TABLE ACCESS FULL         |<\/span> DEPARTMENTS       |<\/span>     1<\/span> |<\/span>    16<\/span> |<\/span>     3<\/span>   (<\/span>0<\/span>)<\/span>|<\/span> 00<\/span>:<\/span>00<\/span>:<\/span>01<\/span> |<\/span>
\n|<\/span>   4<\/span> |<\/span>    INDEX RANGE<\/span><\/a> SCAN          |<\/span> EMP_DEPARTMENT_IX |<\/span>    10<\/span> |<\/span>       |<\/span>     0<\/span>   (<\/span>0<\/span>)<\/span>|<\/span>          |<\/span>
\n|<\/span>   5<\/span> |<\/span>   TABLE ACCESS BY INDEX ROWID|<\/span> EMPLOYEES         |<\/span>    10<\/span> |<\/span>   180<\/span> |<\/span>     1<\/span>   (<\/span>0<\/span>)<\/span>|<\/span> 00<\/span>:<\/span>00<\/span>:<\/span>01<\/span> |<\/span>
\n--------------------------------------------------------------------------------------------------<\/span>
\n
\nColumn Projection Information (<\/span>identified by operation id)<\/span>:<\/span>
\n-----------------------------------------------------------<\/span>
\n
\n   1<\/span> -<\/span> "DEPARTMENT_NAME"<\/span>[<\/span>VARCHAR2,<\/span>30<\/span>]<\/span>,<\/span> "FIRST_NAME"<\/span>[<\/span>VARCHAR2,<\/span>20<\/span>]<\/span>,<\/span> "HIRE_DATE"<\/span>[<\/span>DATE<\/span><\/a>,<\/span>7<\/span>]<\/span>
\n   2<\/span> -<\/span> "DEPARTMENT_NAME"<\/span>[<\/span>VARCHAR2,<\/span>30<\/span>]<\/span>,<\/span> "E"<\/span>.<\/span>ROWID[<\/span>ROWID,<\/span>10<\/span>]<\/span>
\n   3<\/span> -<\/span> "D"<\/span>.<\/span>"DEPARTMENT_ID"<\/span>[<\/span>NUMBER,<\/span>22<\/span>]<\/span>,<\/span> "DEPARTMENT_NAME"<\/span>[<\/span>VARCHAR2,<\/span>30<\/span>]<\/span>
\n   4<\/span> -<\/span> "E"<\/span>.<\/span>ROWID[<\/span>ROWID,<\/span>10<\/span>]<\/span>
\n   5<\/span> -<\/span> "FIRST_NAME"<\/span>[<\/span>VARCHAR2,<\/span>20<\/span>]<\/span>,<\/span> "HIRE_DATE"<\/span>[<\/span>DATE<\/span><\/a>,<\/span>7<\/span>]<\/span><\/div><\/td><\/tr><\/tbody><\/table><\/div>\n

 <\/p>\n

Peeked Binds Section<\/span><\/h2>\n

Quando estamos analisando o desempenho de uma instru\u00e7\u00e3o SQL \u00e9 muito \u00fatil conseguir criar um “TEST CASE” reproduzindo o mesmo plano de execu\u00e7\u00e3o que esta sendo utilizado em produ\u00e7\u00e3o, nesse momento \u00e9 muito importante identificar quais s\u00e3o as vari\u00e1veis que est\u00e3o sendo passadas para a instru\u00e7\u00e3o SQL. A se\u00e7\u00e3o “Peeked Binds” fornece essa informa\u00e7\u00e3o, veja a seguir um exemplo dessa se\u00e7\u00e3o.<\/span><\/p>\n

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
<\/div><\/td>
variable dname varchar2(<\/span>30<\/span>)<\/span>
\nexec<\/span><\/a> :<\/span>dname :=<\/span> 'IT'<\/span>
\n
\nSELECT first_name,<\/span> hire_date,<\/span> department_name
\n  FROM employees emp,<\/span>
\n       departments dpt
\n WHERE emp.<\/span>department_id =<\/span> dpt.<\/span>department_id
\n   AND department_name =<\/span> :<\/span>dname;<\/span>
\n
\nSELECT *<\/span> FROM TABLE(<\/span>dbms_xplan.<\/span>display_cursor(<\/span>'2rkdknf03nj05'<\/span>,<\/span>0<\/span>,<\/span>'TYPICAL LAST +PEEKED_BINDS'<\/span>)<\/span>)<\/span>
\n
\n--------------------------------------------------------------------------------------------------<\/span>
\n|<\/span> Id  |<\/span> Operation                    |<\/span> Name              |<\/span> Rows  |<\/span> Bytes |<\/span> Cost (<\/span>%<\/span>CPU)<\/span>|<\/span> Time<\/span><\/a>     |<\/span>
\n--------------------------------------------------------------------------------------------------<\/span>
\n|<\/span>   0<\/span> |<\/span> SELECT STATEMENT             |<\/span>                   |<\/span>       |<\/span>       |<\/span>     4<\/span> (<\/span>100<\/span>)<\/span>|<\/span>          |<\/span>
\n|<\/span>   1<\/span> |<\/span>  NESTED LOOPS                |<\/span>                   |<\/span>    10<\/span> |<\/span>   340<\/span> |<\/span>     4<\/span>   (<\/span>0<\/span>)<\/span>|<\/span> 00<\/span>:<\/span>00<\/span>:<\/span>01<\/span> |<\/span>
\n|<\/span>   2<\/span> |<\/span>   NESTED LOOPS               |<\/span>                   |<\/span>    10<\/span> |<\/span>   340<\/span> |<\/span>     4<\/span>   (<\/span>0<\/span>)<\/span>|<\/span> 00<\/span>:<\/span>00<\/span>:<\/span>01<\/span> |<\/span>
\n|*<\/span>  3<\/span> |<\/span>    TABLE ACCESS FULL         |<\/span> DEPARTMENTS       |<\/span>     1<\/span> |<\/span>    16<\/span> |<\/span>     3<\/span>   (<\/span>0<\/span>)<\/span>|<\/span> 00<\/span>:<\/span>00<\/span>:<\/span>01<\/span> |<\/span>
\n|*<\/span>  4<\/span> |<\/span>    INDEX RANGE<\/span><\/a> SCAN          |<\/span> EMP_DEPARTMENT_IX |<\/span>    10<\/span> |<\/span>       |<\/span>     0<\/span>   (<\/span>0<\/span>)<\/span>|<\/span>          |<\/span>
\n|<\/span>   5<\/span> |<\/span>   TABLE ACCESS BY INDEX ROWID|<\/span> EMPLOYEES         |<\/span>    10<\/span> |<\/span>   180<\/span> |<\/span>     1<\/span>   (<\/span>0<\/span>)<\/span>|<\/span> 00<\/span>:<\/span>00<\/span>:<\/span>01<\/span> |<\/span>
\n--------------------------------------------------------------------------------------------------<\/span>
\n
\nPeeked Binds (<\/span>identified by position)<\/span>:<\/span>
\n--------------------------------------<\/span>
\n
\n   1<\/span> -<\/span> :<\/span>DNAME (<\/span>VARCHAR2(<\/span>30<\/span>)<\/span>,<\/span> CSID=<\/span>178<\/span>)<\/span>:<\/span> 'IT'<\/span><\/div><\/td><\/tr><\/tbody><\/table><\/div>\n

Note Section<\/span><\/h2>\n

A se\u00e7\u00e3o “Note” fornece informa\u00e7\u00f5es relevantes que foram aplicadas ao plano de execu\u00e7\u00e3o, por exemplo a seguir temos um plano de execu\u00e7\u00e3o que utilizou uma profile durante sua execu\u00e7\u00e3o.<\/span><\/p>\n

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
<\/div><\/td>
select o.<\/span>owner,<\/span> o.<\/span>object_name,<\/span> s.<\/span>segment_type
\nfrom my_segments s,<\/span>
\n     my_objects o
\nwhere o.<\/span>object_name =<\/span> s.<\/span>segment_name
\n  and o.<\/span>owner =<\/span> s.<\/span>owner
\n  and segment_name =<\/span> 'DEPARTMENTS'<\/span>;<\/span>
\n
\n
\nSELECT *<\/span> FROM TABLE (<\/span>dbms_xplan.<\/span>display_cursor (<\/span>'g208t609capq4'<\/span>,<\/span>0<\/span>,<\/span>'TYPICAL LAST -PREDICATE'<\/span>)<\/span>)<\/span>;<\/span>
\n
\n----------------------------------------------------------------------------------<\/span>
\n|<\/span> Id  |<\/span> Operation          |<\/span> Name        |<\/span> Rows  |<\/span> Bytes |<\/span> Cost (<\/span>%<\/span>CPU)<\/span>|<\/span> Time<\/span><\/a>     |<\/span>
\n----------------------------------------------------------------------------------<\/span>
\n|<\/span>   0<\/span> |<\/span> SELECT STATEMENT   |<\/span>             |<\/span>       |<\/span>       |<\/span>   555<\/span> (<\/span>100<\/span>)<\/span>|<\/span>          |<\/span>
\n|<\/span>   1<\/span> |<\/span>  HASH<\/span><\/a> JOIN<\/span><\/a>         |<\/span>             |<\/span>     1<\/span> |<\/span>    65<\/span> |<\/span>   555<\/span>   (<\/span>1<\/span>)<\/span>|<\/span> 00<\/span>:<\/span>00<\/span>:<\/span>07<\/span> |<\/span>
\n|<\/span>   2<\/span> |<\/span>   TABLE ACCESS FULL|<\/span> MY_SEGMENTS |<\/span>     1<\/span> |<\/span>    35<\/span> |<\/span>    83<\/span>   (<\/span>0<\/span>)<\/span>|<\/span> 00<\/span>:<\/span>00<\/span>:<\/span>01<\/span> |<\/span>
\n|<\/span>   3<\/span> |<\/span>   TABLE ACCESS FULL|<\/span> MY_OBJECTS  |<\/span>     2<\/span> |<\/span>    60<\/span> |<\/span>   472<\/span>   (<\/span>1<\/span>)<\/span>|<\/span> 00<\/span>:<\/span>00<\/span>:<\/span>06<\/span> |<\/span>
\n----------------------------------------------------------------------------------<\/span>
\n
\nNote
\n-----<\/span>
\n   -<\/span> SQL profile coe_g208t609capq4_1701876989 used for<\/span> this statement
\n
\n
\n
\nSELECT name,<\/span> sql_text FROM dba_sql_profiles;<\/span>
\n
\nNAME                           SQL_TEXT
\n------------------------------<\/span> -------------------------------------------------------<\/span>
\ncoe_g208t609capq4_1701876989
\n                               select o.<\/span>owner,<\/span> o.<\/span>object_name,<\/span> s.<\/span>segment_type
\n                               from my_segments s,<\/span>
\n                               my_objects o
\n                               where o.<\/span>object_name =<\/span> s.<\/span>segment_name
\n                               and o.<\/span>owner =<\/span> s.<\/span>owner
\n                               and segment_name =<\/span> :<\/span>"SYS_B_0"<\/span><\/div><\/td><\/tr><\/tbody><\/table><\/div>\n

Conclus\u00e3o<\/span><\/h2>\n

Este \u00e9 o ultimo de quatro artigos sobre como fazer a leitura de um plano de execu\u00e7\u00e3o gerado pelo pacote DBMS_XPLAN, nesse artigo procuramos mostrar a import\u00e2ncia das se\u00e7\u00f5es n\u00e3o estat\u00edsticas do plano de execu\u00e7\u00e3o e como elas complementam o entendimento do plano de execu\u00e7\u00e3o. O prop\u00f3sito desse artigo foi trazer algumas informa\u00e7\u00f5es de como essas se\u00e7\u00f5es podem ser interpretadas e o seu relacionamento com o plano de execu\u00e7\u00e3o, com a pr\u00e1tica na utiliza\u00e7\u00e3o desse pacote nas tarefas de tuning de instru\u00e7\u00f5es SQL vamos melhorar nossa habilidade em interpretar essas informa\u00e7\u00f5es e em alguns casos vamos resolver o problema de desempenho das instru\u00e7\u00f5es SQL apenas analisando as informa\u00e7\u00f5es dessas se\u00e7\u00f5es.<\/span><\/p>\n

Refer\u00eancias<\/span><\/h2>\n

http:\/\/allthingsoracle.com\/execution-plans-part-7-query-blocks-and-inline-views\/<\/a><\/p>\n

https:\/\/jonathanlewis.wordpress.com\/2009\/07\/01\/distributed-queries\/<\/a><\/p>\n

https:\/\/asktom.oracle.com\/pls\/apex\/f?p=100:11:::NO:RP:P11_QUESTION_ID:7807480400346035212<\/a><\/p>\n

https:\/\/alexanderanokhin.wordpress.com\/2012\/07\/18\/dont-forget-about-column-projection\/<\/a><\/p>\n

https:\/\/blogs.oracle.com\/optimizer\/entry\/displaying_and_reading_the_execution_plans_for_a_sql_statement<\/a><\/p>\n

 
\n 
\n\"Promo-D75L\"<\/a><\/p>\n

\"video-02\"<\/a>
\n 
\n 
\n 
\n 
\n 
\n 
\n 
\n 
\n 
\n 
\n 
\n 
\n 
\n <\/p>\n","protected":false},"excerpt":{"rendered":"

Nas publica\u00e7\u00f5es anteriores foram abordados os t\u00f3picos: como determinar a sequ\u00eancia de execu\u00e7\u00e3o do plano, como interpretar a estat\u00edsticas estimadas e as estat\u00edsticas geradas durante a execu\u00e7\u00e3o do plano, agora vamos abordar algumas se\u00e7\u00f5es complementares que fornecem informa\u00e7\u00f5es importantes para entender por que o plano de execu\u00e7\u00e3o foi criado com determinadas opera\u00e7\u00f5es, por que essas opera\u00e7\u00f5es utilizam determinados recursos e<\/p>\n","protected":false},"author":2,"featured_media":192,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[35],"tags":[6,54,42,46,36,56,50,10,51,55,29,11,53,49,52,45],"class_list":["post-191","post","type-post","status-publish","format-standard","has-post-thumbnail","hentry","category-execution-plan","tag-cbo","tag-column-projection-information-section","tag-cost-base-optimizer","tag-dbms_xplan-display_cursor","tag-execution-plan","tag-note-section","tag-object-alias","tag-otimizador","tag-outline-data-section","tag-peeked-binds-section","tag-performance-tuning","tag-plano-de-execucao","tag-predicate-information-section","tag-query-block-name-section","tag-remote-section","tag-sql-statement-analysis"],"aioseo_notices":[],"_links":{"self":[{"href":"http:\/\/dbtimewizard.com.br\/blog\/wp-json\/wp\/v2\/posts\/191","targetHints":{"allow":["GET"]}}],"collection":[{"href":"http:\/\/dbtimewizard.com.br\/blog\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"http:\/\/dbtimewizard.com.br\/blog\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"http:\/\/dbtimewizard.com.br\/blog\/wp-json\/wp\/v2\/users\/2"}],"replies":[{"embeddable":true,"href":"http:\/\/dbtimewizard.com.br\/blog\/wp-json\/wp\/v2\/comments?post=191"}],"version-history":[{"count":17,"href":"http:\/\/dbtimewizard.com.br\/blog\/wp-json\/wp\/v2\/posts\/191\/revisions"}],"predecessor-version":[{"id":2065,"href":"http:\/\/dbtimewizard.com.br\/blog\/wp-json\/wp\/v2\/posts\/191\/revisions\/2065"}],"wp:featuredmedia":[{"embeddable":true,"href":"http:\/\/dbtimewizard.com.br\/blog\/wp-json\/wp\/v2\/media\/192"}],"wp:attachment":[{"href":"http:\/\/dbtimewizard.com.br\/blog\/wp-json\/wp\/v2\/media?parent=191"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"http:\/\/dbtimewizard.com.br\/blog\/wp-json\/wp\/v2\/categories?post=191"},{"taxonomy":"post_tag","embeddable":true,"href":"http:\/\/dbtimewizard.com.br\/blog\/wp-json\/wp\/v2\/tags?post=191"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}