Hi,

I only want transactions with PROD's that have PRODGROUP
= '1' to update my cube.

DimTable PROD:
PROD, PRODGROUP

FactTable TRANSACTIONS:
....,PROD,...,QTY

I use starschema-tables and "Optimize schema".

/Jerome

You can set "Source Table Filter" at Cube Editor. If there are multiple
partitions, you can set filtering by partitions through advanced settings of
partition wizard. There you will be able to set filtering based on just
columns, inner joins between tables or both.

Separately, next is additional information. if you use "Optimize schema",
there are things to be considered usually.

If you need to restrict dimension table rows, in other words, if you depend
on the inner joins, between dimension and fact tables, of the SQL query to
exclude fact table rows that contain the dimension keys and if the
dimensions meet optimization condition, you should not use "Optimize
schema". Because "Optimize schema" removes the joins between dimension and
fact tables. Of course, even in this case, you can reset the joins manually
by modifying "Member Key Column" of the level of the dimension in Cube
Editor.

Ohjoo Kwon.


"Jerome" <anonymous (AT) discussions (DOT) microsoft.com> wrote

So if I optimize schema with all dimensions without PROD
dimension in the cube, then add PROD (without optimize
schema) and set the sourcetable filter to what ?

"PROD.PRODGROUP = '1'" ???

/J

The expression rule is "Table"."Column" <operator> Value. For example,
ProdGroup type is numeric.

"Prod"."ProdGroup" = 1

Then the expression is added to where clause.

I recommend you set the filtering expression through partition wizard not
cube editor.

Ohjoo Kwon



"jerome" <anonymous (AT) discussions (DOT) microsoft.com> wrote

I always only use the partition created by AM. Is it
still better to use partition wizard ? Why ?

/J

When you add partitions someday, "Source Table Filter" property of the Cube
will become invalid.

Ohjoo Kwon


"jerome" <anonymous (AT) discussions (DOT) microsoft.com> wrote