Question

您好，我无法理解具有多个表的join。我正在尝试获取每个表的所有数据（计划，预测，理想，实际）。如果同一数据不在vendor..ideal_labor中，则下面的查询将忽略edw..f_labor_dtl中的数据。例如，如果在第edw..f_labor_dtl天没有来自'2019-02-07'的数据，即使该查询包含vendor..ideal_labor上的数据，该查询也将从'2019-02-07'中省略。如果我只有2个表相互连接（使用left/right join或full outer join），则此问题很容易解决。在这里有多个表，我将3维表用作左联接（SB_Finance.dbo.d_Job_Detail，edw..d_date，edw..d_store）。意味着在'2019-02-07'中总会有edw..d_date天。当我在left join上edw..d_date时，在所有临时表上的表现都像inner join一样如何？

with 
fcst as (
    select
        DOB as business_date,
        unit as store_id,
        JobCode as job_code,
        sum(Hours) as Hours
    from vendor..HotSchedulesOptimalsForecast
    group by dob,unit,JobCode
),
schd as (
    select
        DOB as business_date,
        unit as store_id,
        JobCode as job_code,
        sum(ScheduledMinutes/60.0) as Hours
    from vendor..HotSchedulesScheduled
    group by dob,unit,JobCode   
),
ideal as (
    select
        business_date,
        store_id,
        job_code,
        sum(ideal_hrs) as Hours
    from vendor..ideal_labor
    group by business_date,store_id,job_code    
),
act as (
    select
        business_date,
        store_id,
        job_code,
        sum(ttl_hrs) as Hours
    from edw..f_labor_dtl
    group by business_date,store_id,job_code    
)
select
    d.business_date,
    d.f_year,
    d.f_period,
    d.week_period,
    d.week_year,
    d.week_ending,
    s.store_id,
    s.store_desc,
    s.rpt_concept,
    s.region,
    j.job_code,
    j.job_name,
    j.Cat_OPKPI,
    sum(schd.Hours) as Schd_Hrs,
    sum(fcst.Hours) as Fcst_Hrs,
    sum(act.Hours) as Actl_Hrs,
    sum(ideal.Hours) as Ideal_Hrs
from SB_Finance.dbo.d_Job_Detail j
    left join fcst on j.job_code = fcst.job_code
    left join schd on j.job_code = schd.job_code
    left join ideal on j.job_code = ideal.job_code
    left join act act on j.job_code = act.job_code
    left join edw..d_date d on d.business_date = fcst.business_date and d.business_date = schd.business_date and d.business_date = ideal.business_date and d.business_date = act.business_date
    left join edw..d_store s on s.store_id = fcst.store_id and s.store_id = schd.store_id and s.store_id = ideal.store_id and s.store_id = act.store_id
where
    is_closed = 0
    and f_year = 2019
    and Cat_OPKPI in ('Service','Culinary','Support','Janitorial')
group by
    d.business_date,d.f_year,d.f_period,d.week_period,d.week_year,d.week_ending,s.store_id,s.store_desc,s.rpt_concept,s.region,j.job_code,j.job_name,j.Cat_OPKPI

Answer 1

我在玩join之后找到了答案。我在所有维度表上使用cross join，然后使用left join连接其他数据表。

with 
fcst as (
    select
        DOB as business_date,
        unit as store_id,
        JobCode as job_code,
        sum(Hours) as Hours
    from vendor..HotSchedulesOptimalsForecast 
    group by dob,unit,JobCode
),
schd as (
    select
        DOB as business_date,
        unit as store_id,
        JobCode as job_code,
        sum(ScheduledMinutes/60.0) as Hours
    from vendor..HotSchedulesScheduled
    group by dob,unit,JobCode   
),
ideal as (
    select
        business_date,
        store_id,
        job_code,
        sum(ideal_hrs) as Hours
    from vendor..ideal_labor
    group by business_date,store_id,job_code    
),
act as (
    select
        business_date,
        store_id,
        job_code,
        sum(ttl_hrs) as Hours
    from edw..f_labor_dtl
    group by business_date,store_id,job_code    
), summary as (
select
    d.business_date,
    d.f_year,
    d.f_period,
    d.week_period,
    d.week_year,
    d.week_ending,
    s.store_id,
    s.store_desc,
    s.rpt_concept,
    s.region,
    j.job_code,
    j.job_name,
    j.Cat_OPKPI,
    isnull(sum(schd.Hours),0) as Schd_Hrs,
    isnull(sum(fcst.Hours),0) as Fcst_Hrs,
    isnull(sum(act.Hours),0) as Actl_Hrs,
    isnull(sum(ideal.Hours),0) as Ideal_Hrs
from SB_Finance.dbo.d_Job_Detail j
    cross join edw..d_date d
    cross join edw..d_store s
    left join fcst on j.job_code = fcst.job_code and d.business_date = fcst.business_date and s.store_id = fcst.store_id
    left join schd on j.job_code = schd.job_code and d.business_date = schd.business_date and s.store_id = schd.store_id
    left join ideal on j.job_code = ideal.job_code and d.business_date = ideal.business_date and s.store_id = ideal.store_id
    left join act on j.job_code = act.job_code and d.business_date = act.business_date and s.store_id = act.store_id

where
    is_closed = 0
    and d.f_year = 2019
    and Cat_OPKPI in ('Service','Culinary','Support','Janitorial')
group by
    d.business_date,d.f_year,d.f_period,d.week_period,d.week_year,d.week_ending,s.store_id,s.store_desc,s.rpt_concept,s.region,j.job_code,j.job_name,j.Cat_OPKPI

用左联接/外联接联接多个表

1 个答案: