Comments on: Using the OBIEE Aggregate Persistence Wizard

By: LearnerogOBIEE

LearnerogOBIEE — Sun, 25 May 2008 21:40:03 +0000

I am getting this error when i try to run the script from the cmd.

create aggregates
create aggregates
[901][State: S1000] [Oracle][ODBC][Ora]ORA-00901: invalid CREATE command

Statement execute failed

“ag_Sales_Fact”
for “SH”.”Sales Fact”(”Gross Profit Physical”)
at levels (”SH”.”TimesDim”.”Fiscal Month”)
using connection pool “sh”.”Connection Pool”
in “sh”..”SH”
“ag_Sales_Fact”
for “SH”.”Sales Fact”(”Gross Profit Physical”)
at levels (”SH”.”TimesDim”.”Fiscal Month”)
using connection pool “sh”.”Connection Pool”
in “sh”..”SH”
[900][State: 37000] [Oracle][ODBC][Ora]ORA-00900: invalid SQL statement

Statement execute failed

Processed: 2 queries
Encountered 2 errors

By: Marija

Marija — Fri, 16 Nov 2007 10:24:57 +0000

Hi,
I tried to make use of the Aggregate Persistence Wizard but I did not manage to execute the script neither through the NQCMD and neithe through job manager. I got the following error from job manager :

[Oracle][ODBC][Ora]ORA-01017: invalid username/password; logon denied

Connect open failed

From the command line I got a different error:

create aggregates
“ag_F_INVENTORY_CON”
for “biwh”.”F_INVENTORY_CONTROL”(”STORQTY”)
at levels (”biwh”.”Inventory Status”.”InventoryStatus”)
using connection pool “biwh”.”Connection Pool”
in “biwh”.”STAGING_AREA”
[901][State: S1000] [Oracle][ODBC][Ora]ORA-00901: invalid CREATE command

Statement execute failed

Any idea please?

Tnx
Marija

By: Christopher Bridge

Christopher Bridge — Thu, 15 Nov 2007 17:48:29 +0000

Here’s a (Oracle SQL) piece of SQL I find useful to check/review the generated database objects;

select * from all_objects
where 1=1
– 10.1.3.x OBIEE generated object naming
and (
object_name like ‘AG__%’
OR object_name like ‘SA__%’
)
and substr(object_name, 3, 1) = ‘_’
and owner = ‘APPS’
order by timestamp DESC
/

Note the SQL above assumes the ” and owner = ‘APPS’ ”
as that is where I imported the connection pool. I mention this because I found lots of AG_ INDEXES in my eBS (e.g. ENG, EGO schema).

By: Mark Rittman

Mark Rittman — Thu, 01 Nov 2007 13:22:30 +0000

Michael – yes, I suppose you’re right, although this won’t do the sort of incremental updates that we’re more used to with Oracle materialized views. But yes, it could be used as a way of rebuilding the aggregates to a set schedule. Well spotted.

regards

Mark

By: Michael Wilcke

Michael Wilcke — Thu, 01 Nov 2007 10:32:48 +0000

Mark,

in the beginning you write “You can run this script either through the scheduler (the method the docs recommend) or ….”

You seem to have problems to get it working with scheduler, however this could be the solution for regular (i.e. scheduled) updates – right?

Michael

By: Mark Rittman

Mark Rittman — Mon, 29 Oct 2007 13:03:44 +0000

Hi Raymond,

I think the aggregates are meant to be rebuilt each time, i.e. there’s no inbuilt refresh feature. I guess the use case here is that you load an underlying data warehouse, then you use the wizard to create and populate the aggregates, each time – unless anyone else knows of a way to refresh them (apart from running the script each time?)

regards

Mark

By: Raymond de Vries

Raymond de Vries — Mon, 29 Oct 2007 12:01:15 +0000

Fantastic feature, especially if you are not able to use Oracles MV query rewrite option. Up until now, I added self-created aggregate tables manually to the physical layer and in the model layer set the levels. One question I do have is, how are these tables refreshed? The article states that the tables are initially loaded when first created, but how are they kept up to date?