Mysql 更改选择结果的布局

我有下面的查询，我想更改结果的显示方式。我研究过如何使用枢轴，但我不知道如何使用枢轴，如果能得到帮助，我将不胜感激

（缩短的）查询：

我想将数据显示为：

descr       | per_sec   | percent   | offered
under_5_sec | 346       | 91.7772   | 87.3737
under_10_sec| 353       | 93.6340   | 89.1414

有什么建议可以实现吗？

选择
select 
 'under_5_sec' as descr,
 count(if(verb in ('COMPLETEAGENT','COMPLETECALLER') and queue in (9100, 9102, 9104) and data1 <= 5,1, null)) as per_sec,
 count(if(verb in ('COMPLETEAGENT','COMPLETECALLER') and queue in (9100, 9102, 9104) and data1 <= 5,1, null)) / COUNT(if(verb in ('COMPLETEAGENT','COMPLETECALLER') and queue in (9100, 9102, 9104, 9106, 9108, 9600, 9602, 9500, 9502, 9800, 9802),1, null)) * 100 as perct,
 count(if(verb in ('COMPLETEAGENT','COMPLETECALLER') and queue in (9100, 9102, 9104) and data1 <= 5,1, null)) / COUNT(if(verb in ('ENTERQUEUE') and queue in (9100, 9102, 9104, 9106, 9108, 9600, 9602, 9500, 9502, 9800, 9802),1, null)) * 100 as offer,
from queue_log
where partition = 'P001'
and time_id >= unix_timestamp('2016-10-30') and time_id < unix_timestamp('2016-10-31')
UNION
select 
 'under_10_sec' as descr,
 count(if(verb in ('COMPLETEAGENT','COMPLETECALLER') and queue in (9100, 9102, 9104) and data1 <= 10,1, null)) as per_sec,
 count(if(verb in ('COMPLETEAGENT','COMPLETECALLER') and queue in (9100, 9102, 9104) and data1 <= 10,1, null)) / COUNT(if(verb in ('COMPLETEAGENT','COMPLETECALLER') and queue in (9100, 9102, 9104, 9106, 9108, 9600, 9602, 9500, 9502, 9800, 9802),1, null)) * 100 as perct,
 count(if(verb in ('COMPLETEAGENT','COMPLETECALLER') and queue in (9100, 9102, 9104) and data1 <= 10,1, null)) / COUNT(if(verb in ('ENTERQUEUE') and queue in (9100, 9102, 9104, 9106, 9108, 9600, 9602, 9500, 9502, 9800, 9802),1, null)) * 100 as offer
from queue_log
where partition = 'P001'
and time_id >= unix_timestamp('2016-10-30') and time_id < unix_timestamp('2016-10-31')

“5秒以下”作为描述，
count（if）（动词in（'COMPLETEAGENT'，'COMPLETECALLER'）和queue in（9100，9102，9104）以及data1可能重复的感谢，这确实给了我想要的结果。像这样使用union是否运行两次查询（每个union）？这种结果需要联合。我不知道mysql的内部结构。但我认为每个记录对查询求值一次，而不是两次。我的意思是表记录不会一个接一个地与两个查询进行比较。联合创建两个数据集，然后“联合”以形成一个查询结果。感谢T Gray。Stephen您还可以使用union all而不是union来避免两个集合的比较。
descr       | per_sec   | percent   | offered
under_5_sec | 346       | 91.7772   | 87.3737
under_10_sec| 353       | 93.6340   | 89.1414

select 
 'under_5_sec' as descr,
 count(if(verb in ('COMPLETEAGENT','COMPLETECALLER') and queue in (9100, 9102, 9104) and data1 <= 5,1, null)) as per_sec,
 count(if(verb in ('COMPLETEAGENT','COMPLETECALLER') and queue in (9100, 9102, 9104) and data1 <= 5,1, null)) / COUNT(if(verb in ('COMPLETEAGENT','COMPLETECALLER') and queue in (9100, 9102, 9104, 9106, 9108, 9600, 9602, 9500, 9502, 9800, 9802),1, null)) * 100 as perct,
 count(if(verb in ('COMPLETEAGENT','COMPLETECALLER') and queue in (9100, 9102, 9104) and data1 <= 5,1, null)) / COUNT(if(verb in ('ENTERQUEUE') and queue in (9100, 9102, 9104, 9106, 9108, 9600, 9602, 9500, 9502, 9800, 9802),1, null)) * 100 as offer,
from queue_log
where partition = 'P001'
and time_id >= unix_timestamp('2016-10-30') and time_id < unix_timestamp('2016-10-31')
UNION
select 
 'under_10_sec' as descr,
 count(if(verb in ('COMPLETEAGENT','COMPLETECALLER') and queue in (9100, 9102, 9104) and data1 <= 10,1, null)) as per_sec,
 count(if(verb in ('COMPLETEAGENT','COMPLETECALLER') and queue in (9100, 9102, 9104) and data1 <= 10,1, null)) / COUNT(if(verb in ('COMPLETEAGENT','COMPLETECALLER') and queue in (9100, 9102, 9104, 9106, 9108, 9600, 9602, 9500, 9502, 9800, 9802),1, null)) * 100 as perct,
 count(if(verb in ('COMPLETEAGENT','COMPLETECALLER') and queue in (9100, 9102, 9104) and data1 <= 10,1, null)) / COUNT(if(verb in ('ENTERQUEUE') and queue in (9100, 9102, 9104, 9106, 9108, 9600, 9602, 9500, 9502, 9800, 9802),1, null)) * 100 as offer
from queue_log
where partition = 'P001'
and time_id >= unix_timestamp('2016-10-30') and time_id < unix_timestamp('2016-10-31')