Database Reference
In-Depth Information
Relationship
Type
Explanation and examples
Referenced This relationship type indicates that a dimension is related to the
measure group through another dimension. The referenced dimen-
sion and the intermediate dimension must have a common key,
which you specify while defining such a relationship. For example,
the
FactResellerSales
table does not have a
GeographyKey
column, so we cannot use a direct relationship for a geography di-
mension. Instead, we use the Reseller dimension as the intermedi-
ary between
FactResellerSales
and
DimGeography
and spe-
cify
GeographyKey
as the reference dimension attribute and the
intermediate dimension attribute. Note that you have an option to
materialize
the referenced relationship. It is recommended to
materialize
referenced dimensions so that Analysis Services
builds necessary indexes during the processing time instead of
resolving the relationship at query time.
Data mining This relationship type is used for data mining, which is beyond the
scope of this topic.
Many-to-
many
This relationship type requires an intermediary measure group to re-
solve many-to-many relationship between a dimension and a meas-
ure group. For example, each Internet sale could be associated with
multiple sales reasons: television advertisement, manufacturer, re-
view, or promotion. Similarly, each of these reasons for sales is as-
sociated with multiple sales orders. Therefore, you cannot build a
direct relationship between the Internet sales measure group and
the Sales Reason dimension. Instead, you can build an intermedi-
ate fact table called
FactInternetSalesReason
and include
only three columns:
Sales Order Number
,
Sales Order Line
Number
, and
Sales Reason Key
. Next, you can relate the
Search WWH ::
Custom Search