oracle 相同的sql执行两次 执行计划会不一样吗,同样的SQL语句在两个数据库上执行，执行计划差别很大，怎么回事？...

同样的SQL语句在两台数据库服务器上执行(两个数据库的结构都是一样的)，执行计划和响应时间却是相差很大，请问这是怎么回事：

sql语句： select count(*) from table1 a,table1 b,table c

where  c.parentid=b.id and b.parentid=a.id and a.parentid=1 and c.type=5 and b.type=6

/// 执行计划1//

Execution Plan

----------------------------------------------------------

0      SELECT STATEMENT Optimizer=CHOOSE (Cost=76 Card=1 Bytes=24)

1    0   SORT (AGGREGATE)

2    1     HASH JOIN (Cost=76 Card=7 Bytes=168)

3    2       TABLE ACCESS (FULL) OF 'Table1' (Cost=25 Card=511

Bytes=5110)

4    2       MERGE JOIN (CARTESIAN) (Cost=50 Card=3009 Bytes=42126)

5    4         TABLE ACCESS (FULL) OF 'Table1' (Cost=25 Card=1

Bytes=10)

6    4         BUFFER (SORT) (Cost=25 Card=21 Bytes=84)

7    6           TABLE ACCESS (FULL) OF 'Table1' (Cost=25 Card

=21 Bytes=84)

Statistics

----------------------------------------------------------

0  recursive calls

0  db block gets

744  consistent gets

660  physical reads

0  redo size

377  bytes sent via SQL*Net to client

503  bytes received via SQL*Net from client

2  SQL*Net roundtrips to/from client

1  sorts (memory)

0  sorts (disk)

1  rows processed

//执行计划2///

Execution Plan

----------------------------------------------------------

0      SELECT STATEMENT Optimizer=CHOOSE (Cost=198 Card=1 Bytes=23)

1    0   SORT (AGGREGATE)

2    1     NESTED LOOPS (Cost=198 Card=7 Bytes=161)

3    2       NESTED LOOPS (Cost=132 Card=1 Bytes=20)

4    3         TABLE ACCESS (FULL) OF 'Table1' (Cost=66 Card=1

Bytes=10)

5    3         TABLE ACCESS (FULL) OF 'Table1' (Cost=66 Card=1

Bytes=10)

6    2       TABLE ACCESS (FULL) OF 'Table1' (Cost=66 Card=21

Bytes=63)

Statistics

----------------------------------------------------------

0  recursive calls

0  db block gets

6450  consistent gets

4444  physical reads

0  redo size

377  bytes sent via SQL*Net to client

503  bytes received via SQL*Net from client

2  SQL*Net roundtrips to/from client

0  sorts (memory)

0  sorts (disk)

1  rows processed