Deleting requests from Cube with code using infopackage with two targets

Question

hi experts,
Working on BW 3.5.
I am creating a process chain to load from a table in a source system to two Infocubes.
In the start routines of the update rules depending on the date of load, if it is not the last day of month InfoCube A will be loaded with data and InfoCube B will have Datapackage deleted but a request will still load but will contain no data.
And vice versa if it is the last day of month.
In InfoCube A I only want to retain a rolling 15 days of data and have entered a "Delete Overlapping Request from Cube" step.
In the "Deletion-selections" of this step I have added a "Request Selection through Routine" and put in the following code;
  del_date = sy-datum - 15.
  loop at l_t_request_to_delete.
    if l_t_request_to_delete-timestamp+0(8) GT del_date.
      delete l_t_request_to_delete.
    endif.
  endloop.
My question is, how can I ensure this delete code will only work for InfoCube A, as I only want the rollings 15 days of data in here, in Infocube B I want to store all the requests that are loaded here for the end of month.

Answer 1

Hi
Try in some sql browser with outer joins and you will see the problem.
select *
from (select nvl(leverklant.klant_code,'-1') fk_cube_abonnementen_kla_code,
     nvl(factuurklant.klant_code,'-1') fk_cube_abonnementen_kl2_code
     from ods_titsub
     left outer join dim_klant leverklant on (leverklant.klant_code = ods_titsub.cust)
     left outer join dim_klant factuurklant on (factuurklant.klant_code = ods_titsub.sta_cust)
     where (case substr(ods_titsub.isbn,1,8) when 'FBVISTAQ' then 'J' else 'N' end) = 'N'
     and ods_titsub.rec_sta <> 'X'
     and ods_titsub.curr_start >= to_date('01011999', 'DDMMYYYY')
     and ods_titsub.curr_start <= to_date('01022008', 'DDMMYYYY')
     and (leverklant.klant_sk is null or leverklant.klant_sk = leverklant.dimension_key)
     and (factuurklant.klant_sk is null or factuurklant.klant_sk = factuurklant.dimension_key)) a
left outer join dim_klant b on
     b.klant_code = a.fk_cube_abonnementen_kla_code and
     b.dimension_key = b.klant_sk and b.klant_sk is not null
left outer join dim_klant c on
     c.klant_code = a.fk_cube_abonnementen_kl2_code and
     c.dimension_key = c.klant_sk and c.klant_sk is not null
Regards
Result will be like:
FACT_TABLE DIM1 NULL
FACT_TABLE NULL DIM2
but never like:
FACT_TABLE DIM1 DIM2
Message was edited by:
Armands