Question

我有以下查询返回类似＆＃39; top＆＃39; - ＆GT; ＆＃39; 30＆＃39;，＆＃39; metrictype＆＃39; - ＆GT; ＆＃c; cpu usage＆＃39; （该查询计算20％最重的运行机器的平均CPU / MEM使用率。

SELECT   round(avg(avgcpuusage)) as 'top', metrictype
FROM    
    (  
SELECT (vmstatstemp.avgcpuusage),
    vmstatstemp.vmname,
    vmstatstemp.metrictype,
    @counter := @counter +1 AS counter
    FROM (select @counter:=0) AS initvar,  (select round(avg(vmstats.metricvalue)) as 'avgcpuusage', vmstats.vmname,
    vmstats.metrictype from vmstats   
    join vms on vms.naam = vmstats.vmname
    join klanten on vms.klant = klanten.ID
    where metrictype = 'cpu usage'
    AND SUBSTRING(vmstats.dataEntryTimestamp,4,7)  = date_format(now(),'%m:%Y')
    group by vmname
    order by avgcpuusage desc) vmstatstemp
    ) as temp
where counter < round(0.2* (@counter+1))

如果我将此查询与其自身结合（见下文），我会得到以下结果：

＆＃39;顶部＆＃39; - ＆GT; ＆＃39; 30＆＃39;，＆＃39; metrictype＆＃39; - ＆GT; ＆＃c; cpu usage＆＃39; ＆＃39;顶部＆＃39; - ＆GT; ＆＃39; null＆＃39;，＆＃39; metrictype＆＃39; - ＆GT; ＆＃39;空＆＃39;

    SELECT   round(avg(avgcpuusage)) as 'top', metrictype
FROM    
    (  
SELECT (vmstatstemp.avgcpuusage),
    vmstatstemp.vmname,
    vmstatstemp.metrictype,
    @counter := @counter +1 AS counter
    FROM (select @counter:=0) AS initvar,  (select round(avg(vmstats.metricvalue)) as 'avgcpuusage', vmstats.vmname,
    vmstats.metrictype from vmstats   
    join vms on vms.naam = vmstats.vmname
    join klanten on vms.klant = klanten.ID
    where metrictype = 'cpu usage'
    AND SUBSTRING(vmstats.dataEntryTimestamp,4,7)  = date_format(now(),'%m:%Y')
    group by vmname
    order by avgcpuusage desc) vmstatstemp
    ) as temp
where counter < round(0.2* (@counter+1))


union all


SELECT   round(avg(avgcpuusage)) as 'top', metrictype
FROM    
    (  
SELECT (vmstatstemp.avgcpuusage),
    vmstatstemp.vmname,
    vmstatstemp.metrictype,
    @counter := @counter +1 AS counter
    FROM (select @counter:=0) AS initvar,  (select round(avg(vmstats.metricvalue)) as 'avgcpuusage', vmstats.vmname,
    vmstats.metrictype from vmstats   
    join vms on vms.naam = vmstats.vmname
    join klanten on vms.klant = klanten.ID
    where metrictype = 'cpu usage'
    AND SUBSTRING(vmstats.dataEntryTimestamp,4,7)  = date_format(now(),'%m:%Y')
    group by vmname
    order by avgcpuusage desc) vmstatstemp
    ) as temp
where counter < round(0.2* (@counter+1))

我不明白为什么我的结果是＆＃39;顶部＆＃39; - ＆GT; ＆＃39; 30＆＃39;，＆＃39; metrictype＆＃39; - ＆GT; ＆＃c; cpu usage＆＃39; ＆＃39;顶部＆＃39; - ＆GT; ＆＃39; null＆＃39;，＆＃39; metrictype＆＃39; - ＆GT; ＆＃39;空＆＃39; 代替＆＃39;顶部＆＃39; - ＆GT; ＆＃39; 30＆＃39;，＆＃39; metrictype＆＃39; - ＆GT; ＆＃c; cpu usage＆＃39; ＆＃39;顶部＆＃39; - ＆GT; ＆＃39; 30＆＃39;，＆＃39; metrictype＆＃39; - ＆GT; ＆＃39; cpu usage＆＃39;，任何人都能解释一下吗？（并提出如何获得正确结果的建议？）

Answer 1

正如Gordon Linoff指出的那样：你需要为两个子查询使用不同的变量

联合所有相同的查询两次返回null

1 个答案: