Oracle11新特性——SQL快取結果集（三）

打算寫一系列的文章介紹11g的新特性和變化。

這篇文章繼續介紹SQL快取結果集的特點並簡單介紹DBMS_RESULT_CACHE包。

Oracle11新特性——SQL快取結果集（一）：http://yangtingkun.itpub.net/post/468/391015

Oracle11新特性——SQL快取結果集（二）：http://yangtingkun.itpub.net/post/468/391560

上一篇介紹了RESULT CACHE的一些特點，其中最後提到了使用RESULT CACHE可能會導致查詢結果不正確。下面看一個簡單的例子：

SQL> SET AUTOT ON
SQL> ALTER SESSION SET RESULT_CACHE_MODE = MANUAL;

會話已更改。

SQL> SELECT /*+ FULL(T) */ OBJECT_NAME FROM T WHERE ROWNUM = 1;

OBJECT_NAME
------------------------------
ICOL$

執行計劃
----------------------------------------------------------
Plan hash value: 508354683

---------------------------------------------------------------------------
| Id | Operation | Name | Rows | Bytes | Cost (%CPU)| Time |
---------------------------------------------------------------------------
| 0 | SELECT STATEMENT | | 1 | 25 | 2 (0)| 00:00:01 |
|* 1 | COUNT STOPKEY | | | | | |
| 2 | TABLE ACCESS FULL| T | 1 | 25 | 2 (0)| 00:00:01 |
---------------------------------------------------------------------------

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

1 - filter(ROWNUM=1)

統計資訊
----------------------------------------------------------
1 recursive calls
0 db block gets
4 consistent gets
0 physical reads
0 redo size
342 bytes sent via SQL*Net to client
377 bytes received via SQL*Net from client
2 SQL*Net roundtrips to/from client
0 sorts (memory)
0 sorts (disk)
1 rows processed

SQL> SELECT /*+ INDEX(T) */ OBJECT_NAME FROM T WHERE ROWNUM = 1;

OBJECT_NAME
------------------------------
C_OBJ#

執行計劃
----------------------------------------------------------
Plan hash value: 370698254

------------------------------------------------------------------------------------------------
| Id | Operation | Name | Rows | Bytes | Cost (%CPU)| Time |
------------------------------------------------------------------------------------------------
| 0 | SELECT STATEMENT | | 1 | 25 | 3 (0)| 00:00:01 |
|* 1 | COUNT STOPKEY | | | | | |
| 2 | TABLE ACCESS BY INDEX ROWID| T | 1 | 25 | 3 (0)| 00:00:01 |
| 3 | INDEX FULL SCAN | IND_T_OBJECT_ID | 68324 | | 2 (0)| 00:00:01 |
------------------------------------------------------------------------------------------------

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

1 - filter(ROWNUM=1)

統計資訊
----------------------------------------------------------
1 recursive calls
0 db block gets
3 consistent gets
0 physical reads
0 redo size
343 bytes sent via SQL*Net to client
377 bytes received via SQL*Net from client
2 SQL*Net roundtrips to/from client
0 sorts (memory)
0 sorts (disk)
1 rows processed

先將RESULT_CACHE_MODE設定為MANUAL，觀察全表掃描和索引全掃描兩種查詢的結果。由於執行計劃的不同，雖然語句本身是一樣的，但是返回結果中記錄的順序是不同的。在這個例子中，如果使用全表掃描會返回ICOL$，而採用索引掃描，返回結果為C_OBJ#。

下面看看啟用快取結果集的情況：

SQL> SELECT /*+ RESULT_CACHE FULL(T) */ OBJECT_NAME FROM T WHERE ROWNUM = 1;

OBJECT_NAME
------------------------------
ICOL$

執行計劃
----------------------------------------------------------
Plan hash value: 508354683

--------------------------------------------------------------------------------------------------
| Id | Operation | Name | Rows | Bytes | Cost (%CPU)| Time |
--------------------------------------------------------------------------------------------------
| 0 | SELECT STATEMENT | | 1 | 25 | 2 (0)| 00:00:01 |
| 1 | RESULT CACHE | 668yp49bw4mfnak70m2pww541d | | | | |
|* 2 | COUNT STOPKEY | | | | | |
| 3 | TABLE ACCESS FULL| T | 1 | 25 | 2 (0)| 00:00:01 |
--------------------------------------------------------------------------------------------------

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

2 - filter(ROWNUM=1)

Result Cache Information (identified by operation id):
------------------------------------------------------

1 - column-count=1; dependencies=(YANGTK.T); name="SELECT /*+ RESULT_CACHE FULL(T) */ OBJECT_NAME
= 1"

統計資訊
----------------------------------------------------------
1 recursive calls
0 db block gets
4 consistent gets
0 physical reads
0 redo size
342 bytes sent via SQL*Net to client
377 bytes received via SQL*Net from client
2 SQL*Net roundtrips to/from client
0 sorts (memory)
0 sorts (disk)
1 rows processed

SQL> SELECT /*+ RESULT_CACHE INDEX(T) */ OBJECT_NAME FROM T WHERE ROWNUM = 1;

OBJECT_NAME
------------------------------
ICOL$

執行計劃
----------------------------------------------------------
Plan hash value: 370698254

----------------------------------------------------------------------------------------------------
| Id | Operation | Name | Rows | Bytes | Cost (%CPU)| Time |
----------------------------------------------------------------------------------------------------
| 0 | SELECT STATEMENT | | 1 | 25 | 3 (0)| 00:00:01 |
| 1 | RESULT CACHE | 668yp49bw4mfnak70m2pww541d | | | | |
|* 2 | COUNT STOPKEY | | | | | |
| 3 | TABLE ACCESS BY INDEX ROWID| T | 1 | 25 | 3 (0)| 00:00:01 |
| 4 | INDEX FULL SCAN | IND_T_OBJECT_ID | 68324 | | 2 (0)| 00:00:01 |
----------------------------------------------------------------------------------------------------

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

2 - filter(ROWNUM=1)

Result Cache Information (identified by operation id):
------------------------------------------------------

1 - column-count=1; dependencies=(YANGTK.T); name="SELECT /*+ RESULT_CACHE INDEX(T) */ OBJECT_NAM
M = 1"

統計資訊
----------------------------------------------------------
1 recursive calls
0 db block gets
0 consistent gets
0 physical reads
0 redo size
342 bytes sent via SQL*Net to client
377 bytes received via SQL*Net from client
2 SQL*Net roundtrips to/from client
0 sorts (memory)
0 sorts (disk)
1 rows processed

由於第二個查詢指定使用了RESULT CACHE，雖然Oracle顯示的執行計劃是索引掃描，但是利用的快取結果集是全表掃描生成的，所以第二個查詢得到的結果也和第一個查詢一樣。

對於以前的系統中，試圖利用索引按照順序返回結果集的特點來避免排序的寫法，在利用RESULT CACHE時，很可能會出現問題。

不知道Oracle是否會認為這是一個bug，是否以後會有所改進。

RESULT CACHE不僅對查詢的最終結果生效，查詢的中間結果也是可以使用RESULT CACHE的。

在進行測試之前，先將前面的RESULT CACHE清除掉。Oracle提供了一個DBMS_RESULT_CACHE包來管理RESULT CACHE：

SQL> SET AUTOT OFF
SQL> SELECT COUNT(*) FROM V$RESULT_CACHE_OBJECTS;

COUNT(*)
----------
5

SQL> EXEC DBMS_RESULT_CACHE.FLUSH

PL/SQL 過程已成功完成。

SQL> SELECT COUNT(*) FROM V$RESULT_CACHE_OBJECTS;

COUNT(*)
----------
0

這個包還包括其他一些功能，比如BYPASS可以設定當前例項繞過RESULT CACHE機制。INVALIDATE可以設定某個物件關聯的所有RESULT CACHE失效。STATUS返回結果集狀態。

下面看看Oracle如何對查詢的中間結果使用RESULT CACHE：

SQL> SET AUTOT ON
SQL> CREATE TABLE T1 AS SELECT * FROM DBA_SEGMENTS;

表已建立。

SQL> SELECT /*+ RESULT_CACHE */ OWNER, COUNT(*) NUM FROM T GROUP BY OWNER;

OWNER NUM
------------------------------ ----------
WKSYS 371
MDSYS 1281
YANGTK 13
.
.
.
WMSYS 315
SI_INFORMTN_SCHEMA 8

已選擇22行。

執行計劃
----------------------------------------------------------
Plan hash value: 47235625

--------------------------------------------------------------------------------------------------
| Id | Operation | Name | Rows | Bytes | Cost (%CPU)| Time |
--------------------------------------------------------------------------------------------------
| 0 | SELECT STATEMENT | | 22 | 132 | 287 (3)| 00:00:04 |
| 1 | RESULT CACHE | g02167xk9g68s52fnfssvx00tr | | | | |
| 2 | HASH GROUP BY | | 22 | 132 | 287 (3)| 00:00:04 |
| 3 | TABLE ACCESS FULL| T | 68324 | 400K| 282 (1)| 00:00:04 |
--------------------------------------------------------------------------------------------------

Result Cache Information (identified by operation id):
------------------------------------------------------

1 - column-count=2; dependencies=(YANGTK.T); parameters=(nls); name="SELECT /*+ RESULT_CACHE */ O
ROM T GROUP BY OWNER"

統計資訊
----------------------------------------------------------
1 recursive calls
0 db block gets
1011 consistent gets
0 physical reads
0 redo size
826 bytes sent via SQL*Net to client
388 bytes received via SQL*Net from client
3 SQL*Net roundtrips to/from client
0 sorts (memory)
0 sorts (disk)
22 rows processed

SQL> SELECT /*+ RESULT_CACHE */ OWNER, SUM(BYTES)/1024 K FROM T1 GROUP BY OWNER;

OWNER K
------------------------------ ----------
WKSYS 7616
MDSYS 47744
YANGTK 11392
.
.
.
SYS 854656
WMSYS 7296

已選擇18行。

執行計劃
----------------------------------------------------------
Plan hash value: 136660032

--------------------------------------------------------------------------------------------------
| Id | Operation | Name | Rows | Bytes | Cost (%CPU)| Time |
--------------------------------------------------------------------------------------------------
| 0 | SELECT STATEMENT | | 7146 | 209K| 36 (6)| 00:00:01 |
| 1 | RESULT CACHE | f072b82hjsvsm9yvr91n766xwp | | | | |
| 2 | HASH GROUP BY | | 7146 | 209K| 36 (6)| 00:00:01 |
| 3 | TABLE ACCESS FULL| T1 | 7146 | 209K| 34 (0)| 00:00:01 |
--------------------------------------------------------------------------------------------------

Result Cache Information (identified by operation id):
------------------------------------------------------

1 - column-count=2; dependencies=(YANGTK.T1); parameters=(nls); name="SELECT /*+ RESULT_CACHE */
24 K FROM T1 GROUP BY OWNER"

Note
-----
- dynamic sampling used for this statement

統計資訊
----------------------------------------------------------
68 recursive calls
0 db block gets
183 consistent gets
109 physical reads
0 redo size
773 bytes sent via SQL*Net to client
388 bytes received via SQL*Net from client
3 SQL*Net roundtrips to/from client
0 sorts (memory)
0 sorts (disk)
18 rows processed

SQL> SELECT A.OWNER, NUM, K FROM
2 (SELECT /*+ RESULT_CACHE */ OWNER, COUNT(*) NUM FROM T GROUP BY OWNER) A,
3 (SELECT /*+ RESULT_CACHE */ OWNER, SUM(BYTES)/1024 K FROM T1 GROUP BY OWNER) B
4 WHERE A.OWNER = B.OWNER;

OWNER NUM K
------------------------------ ---------- ----------
WKSYS 371 7616
MDSYS 1281 47744
YANGTK 13 11392
.
.
.
SYS 29743 854656
WMSYS 315 7296

已選擇18行。

執行計劃
----------------------------------------------------------
Plan hash value: 3810504953

----------------------------------------------------------------------------------------------------
| Id | Operation | Name | Rows | Bytes | Cost (%CPU)| Time |
----------------------------------------------------------------------------------------------------
| 0 | SELECT STATEMENT | | 7146 | 418K| 323 (3)| 00:00:04 |
|* 1 | HASH JOIN | | 7146 | 418K| 323 (3)| 00:00:04 |
| 2 | VIEW | | 22 | 660 | 287 (3)| 00:00:04 |
| 3 | RESULT CACHE | g02167xk9g68s52fnfssvx00tr | | | | |
| 4 | HASH GROUP BY | | 22 | 132 | 287 (3)| 00:00:04 |
| 5 | TABLE ACCESS FULL| T | 68324 | 400K| 282 (1)| 00:00:04 |
| 6 | VIEW | | 7146 | 209K| 36 (6)| 00:00:01 |
| 7 | RESULT CACHE | f072b82hjsvsm9yvr91n766xwp | | | | |
| 8 | HASH GROUP BY | | 7146 | 209K| 36 (6)| 00:00:01 |
| 9 | TABLE ACCESS FULL| T1 | 7146 | 209K| 34 (0)| 00:00:01 |
----------------------------------------------------------------------------------------------------

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

1 - access("A"."OWNER"="B"."OWNER")

Result Cache Information (identified by operation id):
------------------------------------------------------

3 - column-count=2; dependencies=(YANGTK.T); parameters=(nls); name="SELECT /*+ RESULT_CACHE */ O
ROM T GROUP BY OWNER"

7 - column-count=2; dependencies=(YANGTK.T1); parameters=(nls); name="SELECT /*+ RESULT_CACHE */
24 K FROM T1 GROUP BY OWNER"

Note
-----
- dynamic sampling used for this statement

統計資訊
----------------------------------------------------------
127 recursive calls
8 db block gets
194 consistent gets
0 physical reads
0 redo size
889 bytes sent via SQL*Net to client
388 bytes received via SQL*Net from client
3 SQL*Net roundtrips to/from client
10 sorts (memory)
0 sorts (disk)
18 rows processed

需要注意，對於SQL中間結果使用RESULT CACHE必須使用RESULT_CACHE提示進行強制。

RESULT CACHE功能對於下列情況是無效的：系統表和臨時表；序列的NEXTVAL和CURRVAL偽列；SYSDATE、SYSTIMESTAMP等函式；所有非確定性PL/SQL函式。