注意事项：

1、by keeping them "independent" and then aggregating their aggregates -- we significantly  decreased the amount of processing going on.

在写统计个数的SQL语句的过程，我们最好，先把每个union all的子集的结果进行分别的汇总，然后，最终统一进行汇总，

例如下面两个语句的写法：

A:

SELECT COUNT(*)
  FROM (SELECT NULL
          FROM test_explan
        UNION ALL
        SELECT NULL FROM test_explan);

B:

SELECT SUM(cnt)
  FROM (SELECT COUNT(*) cnt
          FROM test_explan
        UNION ALL
        SELECT COUNT(*) cnt FROM test_explan)

推荐使用第二种写法，参考网址：

http://asktom.oracle.com/pls/asktom/f?p=100:11:0::::P11_QUESTION_ID:21547067945222

和http://blog.csdn.net/tianlesoftware/article/details/5632003

下面是使用跟踪得来的详情：

A

B: