Load Prune

Question

Hi All,
I am getting following error when I try to load into a cube. One of the dimensions of this cube was mapped to three levels from facts table. Assume DIM1 has levels LVL1 , LVL2, LVL3. I load data at all three levels so I mapped same column in facts table to all thee levels. I suspect that is causing the issue. Is there any workaround for this problem?
<ERROR>
<![CDATA[
XOQ-01744: The LOAD PRUNE command cannot be used with cube "TEST.TEST_CUBE".]]>>
</ERROR>
Build Script: TEST.TEST_CUBE USING (LOAD PRUNE, SOLVE)
Olap Version: 11.2.0.3.4
Thanks
Dileep.

Answer 1

Dileep,
You have to first figure out (programmatically) what partitions need to be loaded. You can also have something like a control table where you store all the partitions (lets say MONTHs) for which the cube needs to be loaded. Then dynamically generate DBMS_CUBE.BUILD statement and execute it.
Some sample statements are shown below:
/* Load only the selection of data identified by the WHERE clause */
BEGIN
DBMS_CUBE.BUILD(q'!
GLOBAL."TIME",
GLOBAL.CHANNEL,
GLOBAL.CUSTOMER,
GLOBAL.PRODUCT,
GLOBAL.UNITS_CUBE USING
*(LOAD NO SYNCH*
WHERE UNITS_FACT.MONTH_ID LIKE '2006%'
AND UNITS_FACT.SALES > 5000)!');
END;
/* Another example of loading cube data, with WHERE clause */
begin
dbms_cube.build(q'!
price_cost_cubew using ( FOR global.time
     WHERE HIER_ANCESTOR(WITHIN global.time.fiscal LEVEL global.time.fiscal.fiscal_quarter)= '150'
                                        BUILD(CLEAR, LOAD, SOLVE))!',
     parallelism=>0,
     add_dimensions=>false);
end;
/* Limit the build (using FOR clause) to the SALES measure in 2006.
All objects are built using the complete (C) method. */
BEGIN
DBMS_CUBE.BUILD(
script => '
GLOBAL."TIME",
GLOBAL.CHANNEL,
GLOBAL.CUSTOMER,
GLOBAL.PRODUCT,
GLOBAL.UNITS_CUBE USING
FOR MEASURES(GLOBAL.UNITS_CUBE.SALES)
BUILD(LOAD NO SYNCH WHERE GLOBAL.UNITS_FACT.MONTH_ID LIKE ''2006%'')
method => 'C',
parallelism => 2);
END;
Edited by: Nasar on Jan 30, 2013 2:40 PM