如何将列中的字符串划分为redshift中的列名

Question

我在下面构建了这个查询，以便获得每条路线的总体积、sla_min 以及 sla_status。

sla_status 是使用 case when 语法来计算的，以获得over sla并meet sla

with data_manifest as (
select no,
       concat(concat(origin,'-'),destination) as route_city, 
       sla_min,
       case
           when status>0 and datediff(day, sla_max_date_time_internal, last_valid_tracking_date_time) > 0 then 'OVER SLA'
           when status=0 and datediff(day, sla_max_date_time_internal, current_date) > 0 then 'OVER SLA' else 'MEET SLA'
        end as status_sla
from data
where trunc(tgltransaksi::date) between ('30 January,2023') and ('9 February,2023')
), data_vol as (
select
    route_city,
    count(distinct no) as volume,
    status_sla,
    sla_min,
from data_manifest
group by route_city, status_sla, sla_min
)

查询结果如下：

route_city     vol      status_sla      sla_min
A - B          20        MEET SLA          2
A - B          40        OVER SLA          2
B - C          30        MEET SLA          1
B - C          30        OVER SLA          1

我的问题是如何将MEET SLA和OVER SLA拆分为列名，以便结构如下所示：

route_city    MEET SLA   OVER SLA   total_vol    sla_min
A - B          20           40         60           2
B - C          30           30         60           1 

我应该如何编写查询以在 redshift 中获得所需的结果？

先感谢您

Answer 1

没有看到您的输入数据，不清楚您到底需要什么，但这是一个镜头。

您需要停止按 status_sla 分组并计算 status_sla 的每个值的数量。

with data_manifest as (
select no,
       concat(concat(origin,'-'),destination) as route_city, 
       sla_min,
       case
           when status>0 and datediff(day, sla_max_date_time_internal, last_valid_tracking_date_time) > 0 then 'OVER SLA'
           when status=0 and datediff(day, sla_max_date_time_internal, current_date) > 0 then 'OVER SLA' else 'MEET SLA'
        end as status_sla
from data
where trunc(tgltransaksi::date) between ('30 January,2023') and ('9 February,2023')
), data_vol as (
select
    route_city,
    count(distinct no) as volume,
    count(distinct decode(status_sla, 'MEET SLA', no, NULL)) as meet_sla,
    count(distinct decode(status_sla, 'OVER SLA', no, NULL)) as over_sla,
    sla_min,
from data_manifest
group by route_city, sla_min
)

还有其他方法可以对边缘情况进行投注。 不知道这些是什么导致了这种最小变化的方法。

以上代码未经测试。