OBIEE 11g - serious problem when rotating columns into "table prompt" area

Question

Hi, I'm hitting an OBIEE 11g issue that's causing huge problems on reports. Essentially, the problem is this - if I dump all the columns of a query to a table, everything looks fine. But if I rotate one of the columns into the "table prompts" area, OBIEE is actually resubmitting a WRONG query. The query format changes from:
WITH
SAWITH0 AS (select sum(T38965.ORIGINAL_BUDGET) as c1,
T12637.ORG_ID_DESC as c2
from
GL_ANALYTICS.DIM_ACCOUNT T12582,
GL_ANALYTICS.DIM_ORG T12637,
GL_ANALYTICS.DIM_ACCOUNTING_PERIOD T12597,
GL_ANALYTICS.FACT_LEDGER T38965 / FACT_LEDGER_ytd */*
where  ( T12582.ACCOUNT_KEY = T38965.ACCOUNT_KEY and T12582.STATEMENT_TYPE_CODE = 'Income Statement' and T12597.ACCOUNTING_PERIOD = 11 and T12597.FISCAL_YEAR = T38965.FISCAL_YEAR and T12597.FISCAL_YEAR = 2012 and T12637.ORG_KEY = T38965.ORG_KEY and T38965.FISCAL_YEAR = 2012 and (T12637.DEPARTMENT_CODE_DESC in ('D0200 - Arts Administration', 'D0206 - Film Studies')) and T12597.ACCOUNTING_PERIOD >= T38965.ACCOUNTING_PERIOD )
group by T12637.ORG_ID_DESC)
select D1.c1 as c1, D1.c2 as c2, D1.c3 as c3 from ( select distinct 0 as c1,
D1.c2 as c2,
D1.c1 as c3
from
SAWITH0 D1
order by c2 ) D1 where rownum <= 65001
and changes to:
WITH
SAWITH0 AS (select distinct T12637.ORG_ID_DESC as c1
from
GL_ANALYTICS.DIM_ACCOUNT T12582,
GL_ANALYTICS.DIM_ORG T12637,
GL_ANALYTICS.DIM_ACCOUNTING_PERIOD T12597,
GL_ANALYTICS.FACT_LEDGER T38927 / FACT_LEDGER_period */*
where  ( T12582.ACCOUNT_KEY = T38927.ACCOUNT_KEY and T12597.ACCOUNTING_PERIOD = T38927.ACCOUNTING_PERIOD and T12597.FISCAL_YEAR = T38927.FISCAL_YEAR and T12582.STATEMENT_TYPE_CODE = 'Income Statement' and T12597.ACCOUNTING_PERIOD = 11 and T12597.FISCAL_YEAR = 2012 and T12637.ORG_KEY = T38927.ORG_KEY and T38927.ACCOUNTING_PERIOD = 11 and T38927.FISCAL_YEAR = 2012 and (T12637.DEPARTMENT_CODE_DESC in ('D0200 - Arts Administration', 'D0206 - Film Studies')) ) ),
SAWITH1 AS (select sum(T38965.ORIGINAL_BUDGET) as c1,
T12637.ORG_ID_DESC as c2
from
GL_ANALYTICS.DIM_ACCOUNT T12582,
GL_ANALYTICS.DIM_ORG T12637,
GL_ANALYTICS.DIM_ACCOUNTING_PERIOD T12597,
GL_ANALYTICS.FACT_LEDGER T38965 / FACT_LEDGER_ytd */*
where  ( T12582.ACCOUNT_KEY = T38965.ACCOUNT_KEY and T12582.STATEMENT_TYPE_CODE = 'Income Statement' and T12597.ACCOUNTING_PERIOD = 11 and T12597.FISCAL_YEAR = T38965.FISCAL_YEAR and T12597.FISCAL_YEAR = 2012 and T12637.ORG_KEY = T38965.ORG_KEY and T38965.FISCAL_YEAR = 2012 and (T12637.DEPARTMENT_CODE_DESC in ('D0200 - Arts Administration', 'D0206 - Film Studies')) and T12597.ACCOUNTING_PERIOD >= T38965.ACCOUNTING_PERIOD )
group by T12637.ORG_ID_DESC),
SAWITH2 AS ((select distinct 0 as c1,
D1.c1 as c2,
cast(NULL as  DOUBLE PRECISION  ) as c3
from
SAWITH0 D1
union all
select distinct 1 as c1,
D1.c2 as c2,
D1.c1 as c3
from
SAWITH1 D1))
select D1.c1 as c1, D1.c2 as c2, D1.c3 as c3 from ( select D1.c1 as c1,
D1.c2 as c2,
D1.c3 as c3
from
SAWITH2 D1
order by c1, c2 ) D1 where rownum <= 65001
I completely do not understand what is going on here. The second query has a ton of problems. First off, it's doing some type of UNION operation - no idea why it's doing that. And the real problem is that it's unioning results from a different fact table, FACT_LEDGER_period instead of FACT_LEDGER_ytd, which is completely screwing up our results.
Long and short, I simply don't understand at all why the query should change just from rotating a column up into the table prompt area.
Any ideas?
Thanks,
Scott

Answer 1

Hi Scott,
I see the query to be valid with each CTE doing the following Jobs.....
1. I understand that this particular query is being generated to get the unique "ORG_ID_DESC " to fill up the table prompts.
SAWITH0 AS (select distinct T12637.ORG_ID_DESC as c1
from
GL_ANALYTICS.DIM_ACCOUNT T12582,
GL_ANALYTICS.DIM_ORG T12637,
GL_ANALYTICS.DIM_ACCOUNTING_PERIOD T12597,
GL_ANALYTICS.FACT_LEDGER T38927 / FACT_LEDGER_period /
where ( T12582.ACCOUNT_KEY = T38927.ACCOUNT_KEY and T12597.ACCOUNTING_PERIOD = T38927.ACCOUNTING_PERIOD and T12597.FISCAL_YEAR = T38927.FISCAL_YEAR and T12582.STATEMENT_TYPE_CODE = 'Income Statement' and T12597.ACCOUNTING_PERIOD = 11 and T12597.FISCAL_YEAR = 2012 and T12637.ORG_KEY = T38927.ORG_KEY and T38927.ACCOUNTING_PERIOD = 11 and T38927.FISCAL_YEAR = 2012 and (T12637.DEPARTMENT_CODE_DESC in ('D0200 - Arts Administration', 'D0206 - Film Studies')) ) )
2. This particular query is to get each prompt value and its corresponding data
SAWITH1 AS (select sum(T38965.ORIGINAL_BUDGET) as c1,
T12637.ORG_ID_DESC as c2
from
GL_ANALYTICS.DIM_ACCOUNT T12582,
GL_ANALYTICS.DIM_ORG T12637,
GL_ANALYTICS.DIM_ACCOUNTING_PERIOD T12597,
GL_ANALYTICS.FACT_LEDGER T38965 / FACT_LEDGER_ytd /
where ( T12582.ACCOUNT_KEY = T38965.ACCOUNT_KEY and T12582.STATEMENT_TYPE_CODE = 'Income Statement' and T12597.ACCOUNTING_PERIOD = 11 and T12597.FISCAL_YEAR = T38965.FISCAL_YEAR and T12597.FISCAL_YEAR = 2012 and T12637.ORG_KEY = T38965.ORG_KEY and T38965.FISCAL_YEAR = 2012 and (T12637.DEPARTMENT_CODE_DESC in ('D0200 - Arts Administration', 'D0206 - Film Studies')) and T12597.ACCOUNTING_PERIOD >= T38965.ACCOUNTING_PERIOD )
group by T12637.ORG_ID_DESC)
3. This particular query, I understand is to get the following data (Sample data) relating the prompts and their corresponding data
column (c1) ;column (c2 ie ORG_ID_DESC); column(C3 ie sum(T38965.ORIGINAL_BUDGET))
0;               ORG_ID_DESC1; NULL
1;               ORG_ID_DESC1;10
0;               ORG_ID_DESC2;NULL
1;               ORG_ID_DESC2;40
0;               ORG_ID_DESC3;NULL
1;               ORG_ID_DESC3;29.8
SAWITH2 AS ((select distinct 0 as c1,
D1.c1 as c2,
cast(NULL as DOUBLE PRECISION ) as c3
from
SAWITH0 D1
union all
select distinct 1 as c1,
D1.c2 as c2,
D1.c1 as c3
from
SAWITH1 D1))
4. The last select statement does nothing but selecting the needed data which later gets arranged in the format as needed by the OBIPS
select D1.c1 as c1, D1.c2 as c2, D1.c3 as c3 from ( select D1.c1 as c1,
D1.c2 as c2,
D1.c3 as c3
from
SAWITH2 D1
order by c1, c2 ) D1 where rownum <= 65001
Now, on the table being chosen for picking up "ORG_ID_DESC" might be pretty much depending on the best source available as you know. If you would like BI Server to pick up any particular source, probably you could try the "LTS Priority Group" way.
Hope I was clear and it helps.
Thank you,
Dhar