一、实验说明：
操作系统：redhat 5.4 x86
数据库:oracle 11g R2
该实验通过建立一张规则表和一张不规则表，对这两张表分别使用全表扫描和索引访问，重点在于说明全表扫描的效率也可以优于索引访问。
二、实验具体步骤：
2.1、创建一张规则表jack_test和一张不规则表echo_test。
1 SQL> create table jack_test(x int,y varchar2(2000)); --创建jack_test表 2 3 Table created. 4 5 SQL> begin 6 2 for i in 1..100000 7 3 loop 8 4 insert into jack_test values(i,rpad(dbms_random.random,75,'*')); 9 5 end loop; 10 6 commit; 11 7 end; 12 8 / 13 14 PL/SQL procedure successfully completed. 15 16 SQL> alter table jack_test add constraint jack_pks primary key(x); 17 18 Table altered. 19 20 SQL> create table echo_test nologging --创建echo_test表，并随机插入数据 21 2 as 22 3 select x,y from jack_test 23 4 order by y 24 5 / 25 26 Table created. 27 SQL> alter table echo_test add constraint echo_pk primary key(x); 28 29 Table altered.
2.2、比较对规则表用全表扫描与索引访问：
1 SQL> set timing on; 2 SQL> set linesize 200; 3 SQL> set autotrace traceonly; 4 SQL> select * from jack_test where x between 20000 and 40000; 5 6 20001 rows selected. 7 8 Elapsed: 00:00:00.23 9 10 Execution Plan 11 ---------------------------------------------------------- 12 Plan hash value: 3796984134 13 14 ------------------------------------------------------------------------------- 15 | Id | Operation | Name | Rows | Bytes | Cost (%CPU)| Time | 16 ------------------------------------------------------------------------------- 17 | 0 | SELECT STATEMENT | | 28379 | 27M| 342 (1)| 00:00:05 | 18 |* 1 | TABLE ACCESS FULL| JACK_TEST | 28379 | 27M| 342 (1)| 00:00:05 | --这里的Cost消耗比较大，还有Bytes也比较大 19 ------------------------------------------------------------------------------- 20 21 Predicate Information (identified by operation id): 22 --------------------------------------------------- 23 24 1 - filter("X">=20000 AND "X"<=40000) 25 26 Note 27 ----- 28 - dynamic sampling used for this statement (level=2) 29 30 31 Statistics 32 ---------------------------------------------------------- 33 9 recursive calls 34 0 db block gets 35 2576 consistent gets 36 9 physical reads 37 0 redo size 38 1813724 bytes sent via SQL*Net to client 39 15082 bytes received via SQL*Net from client 40 1335 SQL*Net roundtrips to/from client 41 0 sorts (memory) 42 0 sorts (disk) 43 20001 rows processed 44 45 SQL> analyze table jack_test compute statistics; 46 47 Table analyzed. 48 49 Elapsed: 00:00:01.47 50 SQL> select /*+ index(jack_test jack_pks) */ * from jack_test where x between 20000 and 40000; 51 52 20001 rows selected. 53 54 Elapsed: 00:00:00.20 55 56 Execution Plan 57 ---------------------------------------------------------- 58 Plan hash value: 674810340 59 60 ----------------------------------------------------------------------------------------- 61 | Id | Operation | Name | Rows | Bytes | Cost (%CPU)| Time | 62 ----------------------------------------------------------------------------------------- 63 | 0 | SELECT STATEMENT | | 20002 | 1543K| 282 (0)| 00:00:04 | 64 | 1 | TABLE ACCESS BY INDEX ROWID| JACK_TEST | 20002 | 1543K| 282 (0)| 00:00:04 | 65 |* 2 | INDEX RANGE SCAN | JACK_PKS | 20002 | | 43 (0)| 00:00:01 | 66 ----------------------------------------------------------------------------------------- 67 68 Predicate Information (identified by operation id): 69 --------------------------------------------------- 70 71 2 - access("X">=20000 AND "X"<=40000) 72 73 74 Statistics 75 ---------------------------------------------------------- 76 1 recursive calls 77 0 db block gets 78 2899 consistent gets 79 0 physical reads 80 0 redo size 81 1893672 bytes sent via SQL*Net to client 82 15082 bytes received via SQL*Net from client 83 1335 SQL*Net roundtrips to/from client 84 0 sorts (memory) 85 0 sorts (disk) 86 20001 rows processed
小结：
对于规则表的查询索引访问效率要高于全表扫描。
2.3、比较不规则表用全表扫描与索引访问：
1 SQL> select /*+ full(echo_test) */ * from echo_test where x between 20000 and 40000; 2 3 20001 rows selected. 4 5 Elapsed: 00:00:00.15 6 7 Execution Plan 8 ---------------------------------------------------------- 9 Plan hash value: 3930265464 10 11 ------------------------------------------------------------------------------- 12 | Id | Operation | Name | Rows | Bytes | Cost (%CPU)| Time | 13 ------------------------------------------------------------------------------- 14 | 0 | SELECT STATEMENT | | 21652 | 20M| 333 (1)| 00:00:04 | 15 |* 1 | TABLE ACCESS FULL| ECHO_TEST | 21652 | 20M| 333 (1)| 00:00:04 | 16 ------------------------------------------------------------------------------- 17 18 Predicate Information (identified by operation id): 19 --------------------------------------------------- 20 21 1 - filter("X">=20000 AND "X"<=40000) 22 23 Note 24 ----- 25 - dynamic sampling used for this statement (level=2) 26 27 28 Statistics 29 ---------------------------------------------------------- 30 246 recursive calls 31 0 db block gets 32 2630 consistent gets 33 9 physical reads 34 0 redo size 35 1813724 bytes sent via SQL*Net to client 36 15082 bytes received via SQL*Net from client 37 1335 SQL*Net roundtrips to/from client 38 6 sorts (memory) 39 0 sorts (disk) 40 20001 rows processed 41 42 SQL> select /*+ index(echo_test echo_pk) */ * from echo_test where x between 20000 and 40000; 43 44 20001 rows selected. 45 46 Elapsed: 00:00:00.17 47 48 Execution Plan 49 ---------------------------------------------------------- 50 Plan hash value: 2911547479 51 52 ----------------------------------------------------------------------------------------- 53 | Id | Operation | Name | Rows | Bytes | Cost (%CPU)| Time | 54 ----------------------------------------------------------------------------------------- 55 | 0 | SELECT STATEMENT | | 21652 | 20M| 19916 (1)| 00:03:59 | 56 | 1 | TABLE ACCESS BY INDEX ROWID| ECHO_TEST | 21652 | 20M| 19916 (1)| 00:03:59 | --这里Cost消耗比较大。 57 |* 2 | INDEX RANGE SCAN | ECHO_PK | 21652 | | 45 (0)| 00:00:01 | 58 ----------------------------------------------------------------------------------------- 59 60 Predicate Information (identified by operation id): 61 --------------------------------------------------- 62 63 2 - access("X">=20000 AND "X"<=40000) 64 65 Note 66 ----- 67 - dynamic sampling used for this statement (level=2) 68 69 70 Statistics 71 ---------------------------------------------------------- 72 7 recursive calls 73 0 db block gets 74 21445 consistent gets -- 一致读的数量比较多 75 38 physical reads 76 0 redo size 77 1893672 bytes sent via SQL*Net to client 78 15082 bytes received via SQL*Net from client 79 1335 SQL*Net roundtrips to/from client 80 0 sorts (memory) 81 0 sorts (disk) 82 20001 rows processed
小结：
对于不规则表的查询全表扫描的效率要高于索引访问。其中索引访问中consistent gets高达21445之多，而Cost达到了2w左右。