使用多个case语句合并来自查询的数据

Question

我在数据库中有一些数据，其格式如下：

我编写了以下SQL查询以从该数据库中提取数据。这是我到目前为止的查询：

select network, 
MAX(CASE WHEN channel = 'R' and brand ='B' and spec_ind = 'x' and extended_day_ind = 'y' then 'RB30'
WHEN channel = 'R' and brand ='G' and spec_ind = 'x' and extended_day_ind = 'Y' then 'RG90'
WHEN channel = 'R' and brand ='G' and spec_ind = 'x' and extended_day_ind = 'N' then 'RB30'
WHEN channel = 'R' and brand ='G' and spec_ind = 'x' and extended_day_ind = 'N' then 'RG30' END) as Channel,
case header
    when 'Annual Shift' then value_vl end as AnnualShift,
Case header
    when 'Min' then value_vl end as MinVal,
Case Header
    when 'Max' then value_vl end as MaxVal
From mytable
WHERE network <> 'x'
group by network, value_vl, header
order by network asc

此查询返回如下结果：

这不是很有用，我想像这样合并这些数据：

我将如何解决此问题？谢谢！

Answer 1

我认为您想要条件聚合，其表达方式如下：

select network, 
       (CASE WHEN channel = 'R' and brand ='B' and spec_ind = 'x' and extended_day_ind = 'y' then 'RB30'
                WHEN channel = 'R' and brand ='G' and spec_ind = 'x' and extended_day_ind = 'Y' then 'RG90'
                WHEN channel = 'R' and brand ='G' and spec_ind = 'x' and 
                     extended_day_ind = 'N' then 'RB30'
                WHEN channel = 'R' and brand ='G' and spec_ind = 'x' and extended_day_ind = 'N' then 'RG30'
        END) as Channel,
       MAX(case header when 'Annual Shift' then value_vl end) as AnnualShift,
       MAX(Case header when 'Min' then value_vl end) as MinVal,
       MAX(Case Header  when 'Max' then value_vl end) as MaxVal
From FIN_SANDBOX.asm_paid_rates_range
WHERE network <> 'x'
group by network, 
         (CASE WHEN channel = 'R' and brand ='B' and spec_ind = 'x' and extended_day_ind = 'y' then 'RB30'
              WHEN channel = 'R' and brand ='G' and spec_ind = 'x' and extended_day_ind = 'Y' then 'RG90'
              WHEN channel = 'R' and brand ='G' and spec_ind = 'x' and 
                   extended_day_ind = 'N' then 'RB30'
              WHEN channel = 'R' and brand ='G' and spec_ind = 'x' and extended_day_ind = 'N' then 'RG30'
          END)
order by network asc;

我不记得Teradata是否允许GROUP BY中的列别名。如果这样的话，逻辑可以稍微简化一下。