Question

Sample Data Set and Desired Query Result

我有一份零售店名单，这些零售店随着时间的推移已经改变了所有权（从公司所有，到许可或特许经营，再到公司所有）。我正在尝试跟踪每个所有权类型下每个月开设的商店数量。对于所有权的每次更改，我在数据集中都有一个新的行项目，每个行项目都有相应的更改开始日期和结束日期。

我正在尝试按照每个月的数据按照所有权类型计算商店数量，而且我无法确定如何计算所有月份的商店在所有权开始日期和结束日期之间的数量。希望这张照片清楚地表明了我的目标。

select
    b.fscl_yr_num
    ,b.fscl_per_in_yr_num
    ,a.ownr_type_cd
    ,sum(case when a.line_start_dt < b.end_dt and a.line_end_dt <= b.End_Dt then 1 else 0 end)
from
(
    (
    select *
    from
        (select
            store_num
            ,ownr_type_cd
            ,case when store_term_dt is not null then 'Closed' else 'Open' end as Status
            ,case when to_date(trim(store_open_dt),'DD-MON-YY') > to_date(trim(eff_from_dt),'DD-MON-YY') then to_date(trim(store_open_dt),'DD-MON-YY') else to_date(trim(eff_from_dt),'DD-MON-YY') end as Line_Start_Dt
            ,case when store_term_dt is null then eff_to_dt 
                when to_date(trim(store_term_dt),'DD-MON-YY') < to_date(trim(eff_to_dt),'DD-MON-YY') then to_date(trim(store_term_dt),'DD-MON-YY') else to_date(trim(eff_to_dt),'DD-MON-YY') end as Line_End_Dt
        from
            (select
                store_num
                ,store_open_dt
                ,store_term_dt
                ,eff_from_dt
                ,eff_to_dt
                ,ownr_type_cd
            from
                appca.d_store_vers
            where
                upper(cntry_cd_2_dgt_iso) = 'GB'
                and postal_cd not like ('BT%')
                and store_open_dt is not null
            group by
                store_num
                ,store_open_dt
                ,store_term_dt
                ,eff_from_dt
                ,eff_to_dt
                ,ownr_type_cd
            order by
              store_num
              ,eff_from_dt)
        group by
            store_num
            ,ownr_type_cd
            ,case when store_term_dt is not null then 'Closed' else 'Open' end
            ,case when to_date(trim(store_open_dt),'DD-MON-YY') > to_date(trim(eff_from_dt),'DD-MON-YY') then to_date(trim(store_open_dt),'DD-MON-YY') else to_date(trim(eff_from_dt),'DD-MON-YY') end
            ,case when store_term_dt is null then eff_to_dt 
                when to_date(trim(store_term_dt),'DD-MON-YY') < to_date(trim(eff_to_dt),'DD-MON-YY') then to_date(trim(store_term_dt),'DD-MON-YY') else to_date(trim(eff_to_dt),'DD-MON-YY') end
        order by
            1 asc
            ,2 asc
            ,3 asc)
    where
        to_date(trim(line_start_dt),'DD-MON-YY') < to_date(trim(line_end_dt),'DD-MON-YY')
    ) A
right join

    --Calendar Table--
    (
    select 
        fscl_yr_num, fscl_per_in_yr_num, Cal_dt min(to_date(trim(cal_dt),'DD-MON-YY')) as Start_Dt, max(to_date(trim(cal_dt),'DD-MON-YY')) as End_Dt
    from
        appca.d_cal
    where
        fscl_yr_num is between 1990 and 2018
    group by
        fscl_yr_num, fscl_per_in_yr_num 
    order by 
        1 asc, 2 asc
    ) B
on A.line_end_dt = B.cal_dt
)
group by
    b.fscl_yr_num
    ,b.fscl_per_in_yr_num
    ,a.ownr_type_cd
order by
    b.fscl_yr_num
    ,b.fscl_per_in_yr_num
;

Answer 1

试试这个案例：

sum(case when createTime >(cast(year(createTime) as varchar) +'-'+cast( MONTH(createTime) as varchar)+'-1') and createTime <(cast(year(createTime) as varchar) +'-'+cast( MONTH(createTime) as varchar)+'-31')
then 1 else 0 end) 'Company'

跟踪零售店所有权随时间的变化

1 个答案: