Databases Reference
In-Depth Information
new dimension data to an existing dimension table. Therefore, SSIS provides
an easy way to load new fact and dimension data into your current cube us-
ing SSIS transforms. The two transforms that help in loading such data are
the partition processing transform and the dimension processing transform.
Many large retail stores still use flat files to extract data from the transactional
systems. Your company probably does the same. Often the columns in the
flat files do not contain clean data. During the staging process you clean the
data and format it with appropriate IDs that match your cube in order to load
the data into your cube. SSIS provides transformations to do lookups, get the
correct ids, clean the data on the fly, and then load the data into your cube. In
the following example you will be working with clean data that needs to be
loaded from a flat file into one of the partitions of the AnalysisSer-
vices2005Tutorial cube:
1. Create a new SSIS package under the SSIS packages folder and
name it PipelineDataLoad.
2. The SSIS task that helps you to read data, perform transforms, and
then push the data into a destination is called the Data Flow task.
Drag and drop the Data Flow task into your SSIS editor as shown in
Figure 16-16
and name it Data Flow Partition Load.
Figure 16-16
3. Double-click the Data Flow task. You will now be in the Data Flow
view. The Toolbox window will show you the SSIS transforms avail-
able for use in the Data Flow view. The data flow transforms are cat-
egorized into three main areas, namely, data flow sources, data flow
transformations, and data flow destinations. Data to be loaded into
the partition of the Analysis Services 2005 Tutorial cube is provided
as a flat file in the
Chapter 16
Samples folder named Adventure
Search WWH ::
Custom Search