Delete data for a single dimension member

Question

Is there anyway to delete all the data for a single dimension member from a particular cube?
olap version: 11.2.0.3

Answer 1

Nasar's approach should work, so I would use it so long as it performs well. But I have (several years ago) seen performance issues with CLEAR VALUES when it is applied to something smaller than an entire partition. So here is another method that also uses the FOR...BUILD construct. It clears the value of the PRICE measure in PRICE_COST_CUBE for PRODUCT = '48'. The cube has only two dimensions, TIME and PRODUCT, and I only want to assign to the leaf level of TIME, which is MONTH.
BEGIN
  DBMS_CUBE.BUILD(Q'!
   "PRICE_COST_CUBE" USING(
    FOR
     "TIME" LEVELS ("TIME"."MONTH"),
     "PRODUCT" WHERE "PRODUCT".DIM_KEY = '48'
    BUILD (SET PRICE_COST_CUBE.PRICE = NULL ))!',
    PARALLELISM=>0,
    ADD_DIMENSIONS=>FALSE);
END;
/You can increase parallelism here if your cube is partitioned. If you need to clear out other measures at the same time, then add additional SET command in the central block. E.g.
BUILD
  SET PRICE_COST_CUBE.PRICE = NULL,
  SET PRICE_COST_CUBE.COST = NULL
)After this step you can re-aggregate the cube to make the aggregate values reflect the change.
BEGIN
  DBMS_CUBE.BUILD(Q'!
   "PRICE_COST_CUBE" USING(
    FOR
     "TIME" LEVELS ("TIME"."MONTH"),
     "PRODUCT" WHERE "PRODUCT".DIM_KEY = '48'
    BUILD (SET PRICE_COST_CUBE.PRICE = NULL ),
    SOLVE)!',
    PARALLELISM=>0,
    ADD_DIMENSIONS=>FALSE);
END;
/The simple assignment to NULL will generate a simple assignment in the OLAP DML.
PRICE_COST_CUBE_STORED=NAYou can force the code to loop over the existing composite by replacing NULL with something that evaluates to NULL, but involves the composite.
SET PRICE_COST_CUBEW.PRICE = NVL2(NULL, PRICE_COST_CUBE.PRICE, NULL)Here is the generated OLAP DML assignment.
PRICE_COST_CUBEW_STORED=NVL2(NA, GLOBAL.GLOBAL!PRICE_COST_CUBEW_PRICE
, NA)  ACROSS GLOBAL.GLOBAL!PRICE_COST_CUBEW_COMPOSITE