我有来自下面查询的数据，看起来像下面的第一个表，如果您观察结果是基于count desc。我想以第二个表的形式显示数据，我遇到了查询问题。也就是说，如果pagetype是详细信息，那么我想要对该ID的计数求和，如果pagetype是singe-item，我想不管它，并按所有结果排序desc。 第一个查询要点如下所示，我还有很多其他内容，但这里有一个简化版本。
SELECT id, title, count(id) as count_num , pagetype
, ROW_NUMBER() OVER ( ORDER BY count(id) desc ) AS the_row
FROM table1, table2
where pagetype in ('details','items','single-item')
and table1.id = table2.id
AND ct.PAGE_VIEW_DT > sysdate - 90
ORDER BY the_row
ID Title Count pagetype
--------------------------------------------------------------
33969 TITLE ONE 523 details
33969 TITLE ONE 494 items
198068 TITLE THREE 400 single-item
33968 TITLE TWO 395 details
198068 TITLE THREE 391 items
198068 TITLE THREE 333 items
198068 TITLE THREE 281 details
33969 TITLE ONE 280 single-item
33968 TITLE TWO 270 single-item
ID Title Count pagetype
--------------------------------------------------------------
33969 TITLE ONE 1017 details,items
198068 TITLE THREE 1005 details,items
198068 TITLE THREE 400 single-item
33968 TITLE TWO 395 details
33969 TITLE ONE 280 single-item
33968 TITLE TWO 270 single-item
答案 0 :(得分：1)
考虑：条件聚合。由于我在这里添加了一个组ID，Title将事物分开，我将pagetype中的值更改为详细信息，项目遇到详细信息或项目时，从而允许group by工作。我认为嵌套替换也可以使group by按照需要工作。
SELECT id, title, count(id) as count_num
, case when pagetype in ('details','items')
then 'details,items'
else pagetype end as pagetype
-- , ROW_NUMBER() OVER ( ORDER BY count(id) desc ) AS the_row
FROM table1, table2
where pagetype in ('details','items','single-item')
AND ct.PAGE_VIEW_DT > sysdate - 90
GROUP BY ID, Title, case when pagetype in ('details','items')
then 'details,items'
else pagetype end
ORDER BY --the_row
count(ID) DESC;
---重读和重读问题后尝试两个...... 感谢@АнатолийПредеин为with语句节省了一些时间......
with a(ID,Title,Count,pagetype) as (
-- here your data from question
select 33969, 'TITLE ONE', 523, 'details' from dual union all
select 33969, 'TITLE ONE', 494, 'items' from dual union all
select 198068, 'TITLE THREE', 400,'single-item' from dual union all
select 33968, 'TITLE TWO', 395,'details' from dual union all
select 198068, 'TITLE THREE', 391, 'items' from dual union all
select 198068, 'TITLE THREE', 333,'items' from dual union all
select 198068, 'TITLE THREE', 281,'details' from dual union all
select 33969, 'TITLE ONE', 280,'single-item' from dual union all
select 33968, 'TITLE TWO', 270,'single-item' from dual ),
with a(ID,Title,Count,pagetype) as (
-- here your data from question
select 33969, 'TITLE ONE', 523, 'details' from dual union all
select 33969, 'TITLE ONE', 494, 'items' from dual union all
select 198068, 'TITLE THREE', 400,'single-item' from dual union all
select 33968, 'TITLE TWO', 395,'details' from dual union all
select 198068, 'TITLE THREE', 391, 'items' from dual union all
select 198068, 'TITLE THREE', 333,'items' from dual union all
select 198068, 'TITLE THREE', 281,'details' from dual union all
select 33969, 'TITLE ONE', 280,'single-item' from dual union all
select 33968, 'TITLE TWO', 270,'single-item' from dual )
Select ID, Title
, sum(count)
, case when pageType in ('details','items') then 'details,items' else pagetype end as PageType
FROM a
Group by ID, Title , case when pageType in ('details','items') then 'details,items' else pagetype end
order by sum(count) desc
在初始查询中使用条件聚合可能有一种更简单的方法...
类似
SELECT id
, title
, count(id) as count_num
, case when pagetype in ('details','items')
then 'details,items'
else pagetype end as pagetype,
FROM table1
INNER JOIN table2
on table1.id = table2.id
WHERE pagetype in ('details','items','single-item')
AND ct.PAGE_VIEW_DT > sysdate - 90
GROUP BY ID
, Title
, case when pagetype in ('details','items')
then 'details,items'
else pagetype end
ORDER BY count(id) desc
但如果没有问题中未提供的原始源数据，我无法测试它。 （而且我不想嘲笑一些！）
答案 1 :(得分：0)
您可以像这样使用双listagg和oracle decode以及oracle with a(ID,Title,Count,pagetype) as (
-- here your data from question
select 33969, 'TITLE ONE', 523, 'details' from dual union all
select 33969, 'TITLE ONE', 494, 'items' from dual union all
select 198068, 'TITLE THREE', 400,'single-item' from dual union all
select 33968, 'TITLE TWO', 395,'details' from dual union all
select 198068, 'TITLE THREE', 391, 'items' from dual union all
select 198068, 'TITLE THREE', 333,'items' from dual union all
select 198068, 'TITLE THREE', 281,'details' from dual union all
select 33969, 'TITLE ONE', 280,'single-item' from dual union all
select 33968, 'TITLE TWO', 270,'single-item' from dual )
select id, title, sum(scount) scount, listagg(pagetype,',') WITHIN group (order by pagetype)
from (
select id, title, sum(count) scount, pagetype
from a
group by id, title, pagetype
)
group by id, title, decode(pagetype,'details','tosum','items','tosum',pagetype)
order by scount desc
函数（此查询基于您查询中的数据）：
33969 TITLE ONE 1017 details,items
198068 TITLE THREE 1005 details,items
198068 TITLE THREE 400 single-item
33968 TITLE TWO 395 details
33969 TITLE ONE 280 single-item
33968 TITLE TWO 270 single-item
此查询的输出：
with a as (
SELECT id, title, count(id) as count_num , pagetype
, ROW_NUMBER() OVER ( ORDER BY count(id) desc ) AS the_row
FROM table1, table2
where pagetype in ('details','items','single-item')
AND ct.PAGE_VIEW_DT > sysdate - 90
)
select id, title, sum(scount) scount, listagg(pagetype,',') WITHIN group (order by pagetype) ptype
from (
select id, title, sum(count_num) scount, pagetype
from a
group by id, title, pagetype
)
group by id, title, decode(pagetype,'details','tosum','items','tosum',pagetype)
order by scount desc
可能看起来像这样
TITLE THREE它的双重分组是因为'details,items,items'中的listagg concatinate与'details,items'重复的字符串而不是你想要group by这就是为什么我使用双ManagementObjectSearcher searcher = new ManagementObjectSearcher("root\\CIMV2",
"SELECT * FROM Win32_PnPEntity");
//This loops through the results from the searcher
foreach (ManagementObject queryObj in searcher.Get())
{
//If it finds the port,
if (queryObj["Caption"].ToString().Contains("##### Wireless AT"))
{
//it writes it to the file
sw.WriteLine("serial port : {0}", queryObj["Caption"] + "\n");
sw.Flush();
}