相同查询条件，新增了关联的表，确提升了查询效率！？求大师
SQL1：
SELECT COUNT(*)
FROM tbm_order_queue_his a, tbm_main_inst_his c
WHERE a.order_id = c.order_id
AND a.state = 1
AND c.is_new = 1
AND a.offer_id IN (SELECT tso.offer_id
FROM tpm_service_offer tso
WHERE tso.offer_type_id = 1)
AND a.prod_id IN (SELECT tmp.main_prod_id
FROM tpm_main_product tmp
WHERE tmp.prod_big_id in (11, 12))
AND c.company_id in
(select tc.company_id
from tsm_company tc
start with tc.company_id = 1668
connect by PRIOR tc.company_id = tc.upper_id)
AND a.finish_time BETWEEN
to_date('2012-02-01 11:35:17', 'yyyy-mm-dd hh24:mi:ss') AND
to_date('2012-02-28 11:35:20', 'yyyy-mm-dd hh24:mi:ss')
SQL2：
SELECT count(*)
FROM tbm_order_queue_his a, tbm_main_inst_his c, tbm_cust_his d
WHERE a.order_id = c.order_id
AND a.order_id = d.order_id
AND a.state = 1
AND c.is_new = 1
AND d.is_new = 1
AND a.offer_id IN (SELECT tso.offer_id
FROM tpm_service_offer tso
WHERE tso.offer_type_id = 1)
AND a.prod_id IN (SELECT tmp.main_prod_id
FROM tpm_main_product tmp
WHERE tmp.prod_big_id in (11, 12))
AND c.company_id in
(select tc.company_id
from tsm_company tc
start with tc.company_id = 1668
connect by PRIOR tc.company_id = tc.upper_id)
AND a.finish_time BETWEEN
to_date('2012-02-01 11:35:17', 'yyyy-mm-dd hh24:mi:ss') AND
to_date('2012-02-28 11:35:20', 'yyyy-mm-dd hh24:mi:ss')
SQL1执行时间：501秒
SQL2执行时间：2秒以内
不明白为何多关联一个表效率反而提高了。
SQL如上所示，查询条件中以下列建了索引（a.offer_id，a.prod_id，c.company_id，a.finish_time，a.order_id，d.order_id）
该时间段数据量10W左右。总数据量300W左右。
若SQL1去掉条件：
AND c.company_id in
(select tc.company_id
from tsm_company tc
start with tc.company_id = 1668
connect by PRIOR tc.company_id = tc.upper_id)
则执行效率也在2秒以内，SQL2有SQL1的所有查询条件，但SQL2为何还会能快？寻大师指点。(SQL2中tbm_cust_his表 order_id建了索引) sql
[解决办法]
3个表里的order_id都是唯一的吗？或者至少两个表里是唯一的？否则出来的结果可能是错误的。。
另外，d表是不是数据量较少？
最好是把两个语句的执行计划都贴出来这才分析得出来……
