Re: Most efficient SQL approach

[Charles Wilt <charles.wilt@xxxxxxxxx>]

A subselect like this is a very bad idea...

better idea

with ord_ytd as (
  select com_nbr, ent_nbr, coalesce(sum(opn_amt),0)
       from V_ORDERS
     where  ord_date>='2020-01-01'
  group by com_nbr, ent_nbr
), inv_ytd as (
  select com_nbr, ent_nbr, coalesce(sum(shp_amt),0)
     from V_INVOICES
  where inv_date>='2020-01-01'
 group by com_nbr, ent_nbr
), <...>
select CUS.COM_NBR, CUS.CUST_NBR, CUS.CUST_NAME, (more here),
from V_CUSTMSTR as CUS
       left outer join ord_ytd  CO on cus.com_nbr = CO.com_nbr and
cus.ent_nbr = CO.ent_nbr
       left outer join inv_ytd  CI on cus.com_nbr = CI.com_nbr and
cus.ent_nbr = CI.ent_nbr
       <...>

The idea is that you build separate results sets, from the various tables
aggregated to the proper level; in this case com_nbr and ent_nbr.  Then
simply join them together.

You could even build summary views (and perhaps supporting EVIs) to use in
place of the CTEs.

You might be interested in the following presentation
https://www.linkedin.com/posts/charles-wilt-2a4a303_sql-for-the-rpg-developer-activity-6641756015354355712-woZR


Charles




On Fri, Mar 6, 2020 at 9:52 AM Greg Wilburn <
gwilburn@xxxxxxxxxxxxxxxxxxxxxxx> wrote:

> I am rewriting an old RPG report using SQL (and SQL views I already
> have).  The "report" generates a database table as output
>
> I need one row for each customer.  In each row I need columns for several
> sales figures that can be found in two other views.
> I have SQL views for the Customer, Invoices and Orders.
>
> I currently have it written where each sales figure is a separate
> sub-select.  Note that the "where" clause for each is slightly different.
>
> It seems like this could be written more efficiently... but I'm not sure
> what approach to take
>
> 1.      Use common table expressions?
>
> 2.      Create the output table, then use separate SQL expressions to
> insert each of the summed values?
>
> This is a simplified version of the code (for simplicity I've replaced the
> RPG variables with constant values).
>
> SELECT CUS.COM_NBR, CUS.CUST_NBR, CUS.CUST_NAME, (more here),
> (select coalesce(sum(opn_amt),0) from V_ORDERS where com_nbr=cus.com_nbr
> and ent_nbr=cus.ent_nbr and
>          ord_date>='2020-01-01' and ord_date<='2020-03-04') as BOOK_YTD,
> (select coalesce(sum(shp_amt),0) from V_INVOICES where com_nbr=cus.com_nbr
> and ent_nbr=cus.ent_nbr and
>          inv_date>='2020-01-01' and inv_date<='2020-03-04' )as NET_YTD,
> (select coalesce(sum(shp_amt),0) from V_INVOICES where com_nbr=cus.com_nbr
> and ent_nbr=cus.ent_nbr and
>          inv_date>='2019-01-01' and inv_date<='2019-03-04' ) as
> PREV_YR_TOTAL,
> (select coalesce(sum(shp_amt),0) from V_INVOICES where com_nbr=cus.com_nbr
> and ent_nbr=cus.ent_nbr and
>          inv_date>='2019-01-01' and inv_date<='2019-03-04' ) as
> NET_PREV_YTD,
> (select coalesce(sum(ord_amt),0) from V_ORDERSA where com_nbr=cus.com_nbr
> and ent_nbr=cus.ent_nbr and
>          ord_date>='2019-01-01' and ord_date<='2019-03-04') as
> BOOK_PREV_YTD
> from V_CUSTMSTR as CUS
> where CUS.COM_NBR='012';
>
> Thanks,
> Greg
> --
> This is the RPG programming on IBM i (RPG400-L) mailing list
> To post a message email: RPG400-L@xxxxxxxxxxxxxxxxxx
> To subscribe, unsubscribe, or change list options,
> visit: https://lists.midrange.com/mailman/listinfo/rpg400-l
> or email: RPG400-L-request@xxxxxxxxxxxxxxxxxx
> Before posting, please take a moment to review the archives
> at https://archive.midrange.com/rpg400-l.
>
> Please contact support@xxxxxxxxxxxx for any subscription related
> questions.
>
> Help support midrange.com by shopping at amazon.com with our affiliate
> link: https://amazon.midrange.com