Question

请帮我解决这个问题。

表是这样的 -

Type | SubType | Flag   | Value
--------------------------------
123 |   A1     |   Y    |  101
--------------------------------
123  |  A2     |   Y    |  102
------------------------------ 
123  |  A3     |   Y    |  103
------------------------------ 
124  |  A4     |   N    |  104
------------------------------ 
124  |  A5     |   N    |  105
------------------------------ 
124  |  A6     |   N    |  106
------------------------------ 
125  |  A7     |   Y    |  107
------------------------------ 
125  |  A8     |   Y    |  108
------------------------------ 
125  |  A9     |   N    |  109
------------------------------ 
125  |  A10    |   N    |  110

要求是根据特定条件选择行 - 如果特定类型的所有标志都是Y，那么只选择具有最低子类型的行，如果特定类型的所有标志都是N，则选择该类型的所有行，如果一个类型的标志是Y和N的组合，那么为该类型选择2行 - 一个具有标志Y的最低子类型，一个具有标志N的最低子类型。

所以上表的输出应该是 -

Type  | SubType | Flag   | Value
------------------------------ 
 123  |  A1     |   Y    |  101
------------------------------ 
 124  |  A4     |   N    |  104
------------------------------ 
 124  |  A5     |   N    |  105
------------------------------ 
 124  |  A6     |   N    |  106
------------------------------ 
 125  |  A7     |   Y    |  107
------------------------------ 
 125  |  A9     |   N    |  109

对于糟糕的表格格式感到抱歉，请提前感谢您对此进行调查。

Answer 1

你可以使用这样的窗口函数：

with your_table (Type , SubType , Flag   , Value) as (
select 123  ,  'A1'     ,   'Y'    ,  101 union all
select 123  ,  'A2'     ,   'Y'    ,  102 union all
select 123  ,  'A3'     ,   'Y'    ,  103 union all
select 124  ,  'A4'     ,   'N'    ,  104 union all
select 124  ,  'A5'     ,   'N'    ,  105 union all
select 124  ,  'A6'     ,   'N'    ,  106 union all
select 125  ,  'A7'     ,   'Y'    ,  107 union all
select 125  ,  'A8'     ,   'Y'    ,  108 union all
select 125  ,  'A9'     ,   'N'    ,  109 union all
select 125  ,  'A10'    ,   'N'    ,  110)

select Type, SubType, Flag, Value
from (
  select
    t.*,
    case when min(flag) over (partition by type)  = max(flag) over (partition by type) then 1 else 2 end cnt,
    row_number() over (partition by type, flag order by convert(int, substring(subtype,2,len(subtype)))) rn
  from your_table t
) t where 
(cnt = 1 and (rn = 1 or flag = 'N'))
or (cnt <> 1 and rn = 1);

产地：

Answer 2

使用CTE为每种类型生成摘要，然后对其进行过滤：

g0 <- [... the rest of the graph ...]

cor.value <- 0.84
c.a <- substitute(~rho == x, list(x=cor.value))
g0 + labs(y="Angekreuzte Option", x=c.a)

在上面的代码中，我们从SubType中提取数字以正确排序。这可确保WITH summary AS ( SELECT [Type] ,SUM(CASE WHEN [Flag] = 'Y' THEN 1 ELSE 0 END) AS Count_Y ,SUM(CASE WHEN [Flag] = 'N' THEN 1 ELSE 0 END) AS Count_N -- This value will be NULL, if there are no rows with Y flag ,MIN(CASE WHEN [Flag] = 'Y' THEN CAST(SUBSTRING(SubType,2,99) AS INT) END) AS Lowest_Y_SubType -- This value will be NULL, if there are no rows with N flag ,MIN(CASE WHEN [Flag] = 'N' THEN CAST(SUBSTRING(SubType,2,99) AS INT) END) AS Lowest_N_SubType FROM dbo.tSO_41657760 GROUP BY [Type] ) SELECT * FROM dbo.tSO_41657760 t WHERE EXISTS ( SELECT 1 FROM summary -- Filter this inner table using Type from the outer table WHERE [Type] = t.[Type] AND ( -- All rows have N flag (Count_Y = 0) -- Select the lowest N and Y rows with lowest SubType values OR (CAST(SUBSTRING(SubType,2,99) AS INT) IN (Lowest_N_SubType, Lowest_Y_SubType)) ) ) 小于 A9（否则，纯字母排序会产生不正确的结果）。

Answer 3

这是一种方式

Group.find({})
     .populate({
         path : 'teachers' , 
         match : { username : "bob" }
      })
      .exec(callback);

Live Demo

根据特定条件筛选行

3 个答案: