The Data Source View
One of the most powerful concepts in an Analysis Services project is that of the Data
Source View, or DSV. A DSV is a logical construct that represents how the structure
of the underlying database in the data source should look to Analysis Services. In its
simplest form, it is merely a copy of the schema from the data source. Developers
can simply add all the tables from the database in the data source and be done. The
database schema is read, the joins between tables are automatically added to the
DSV, and then the cube can be built.
The DSV has far greater capabilities than simply allowing a copy of the
underlying database schema. First, not all the tables have to be brought in, as a
subset will often suffice. This is especially true when building a data mart for a
particular group. If the relational warehouse contains more than that group cares to
see, the DSV would only include the tables relevant to that particular group.
Figure 3-7 shows part of the wizard that steps developers through the process
of creating a DSV, using the Adventure Works DW sample database that ships
with SQL Server 2005. Remember that with Analysis Services 2005, a single cube
can contain multiple fact tables. This helps in situations such as recording sales at
a daily level, but having a forecast or budget table at the month or quarter level.
Figure 3-7 shows that there are two fact tables in this cube: FactSalesQuota and
FactResellerSales. A subset of these dimensions has been added as well, in order to
keep this example simple.
Pages:
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97