Previous Page | Next Page

Overview of SAS/Warehouse Administrator

Task Summaries

Many different approaches can be taken when building a data warehouse. Here are some general guidelines:


Plan Your Data Warehouse

  1. Identify the business problem(s) to be solved.

  2. Identify the subject areas to be included in your data warehouse (such as Sales, Products, Customers, and so on).

  3. Choose a subject area to be developed. The first subject area should be important enough to illustrate the value of your project, but simple enough so that it can be developed quickly and successfully.

  4. Draft the reports and other outputs that you expect to get from the chosen subject area.

  5. Identify data sources required to produce the reports and other outputs that you expect to get from the chosen subject area.

  6. Identify the main column mappings and data transformations between sources and targets in the chosen subject area.

  7. Select the hardware and software required to access the data sources, execute the transformations, and store the refined data in the appropriate formats and locations.

  8. Choose the SAS/Warehouse Administrator data stores that are appropriate for the chosen subject area.

  9. Choose the appropriate exploitation tool(s) for the chosen subject area (such as SAS/EIS for multidimensional analysis; Enterprise Miner for data mining, and so on).

A detailed explanation of how to create a project plan for your data warehouse is beyond the scope of this document. However, the following chapters will help you with the SAS/Warehouse Administrator portion of such a plan:


Create a Data Warehouse Environment

  1. Create a directory structure for the Data Warehouse Environment.

  2. If necessary, update the SAS configuration file and the SAS autoexec file that is used during the SAS/Warehouse Administrator session so that any required options or librefs are available.

  3. Add a Data Warehouse Environment.

    1. Define global metadata for the Environment.

    2. Register data sources for the Environment (create Operational Data Definitions).

      1. Enter metadata for the columns and location of the data sources.

      2. Define one or more Jobs for the data sources.

      3. Execute the Job(s) for the data sources.

      4. Verify that the data sources are available.

For details about the tasks in this section, see the following chapters:

If you need to convert a Data Warehouse Environment created with an earlier version of SAS/Warehouse Administrator, see Converting Metadata for Environments and Warehouses.


Create a Data Warehouse

  1. Add a Data Warehouse within the Environment.

  2. Create a Subject within the Warehouse. Name it after a subject area identified in your project plan.

  3. Create appropriate detail data stores within the Subject.

    1. Enter metadata for the data sources, columns, and locations of the detail data stores.

    2. Define one or more Jobs for the detail data stores.

    3. Execute the Job(s) for the detail data stores.

    4. Verify that the detail data stores are available.

  4. If required by your project plan, create appropriate summary data stores within the Subject.

    1. Enter metadata for the data sources, columns, and locations of the summary data stores.

    2. Define one or more Jobs for the summary data stores.

    3. Execute the Job(s) for the summary data stores.

    4. Verify that the summary data stores are available.

  5. If required by your project plan, create appropriate Information Marts in the Subject.

    1. Enter metadata for the InfoMart Items and InfoMart Files.

    2. Define one or more Jobs for these objects.

    3. Execute the Job(s).

    4. Verify that the InfoMart Items and InfoMart Files are available.

For details about the tasks in this section, see the following chapters:


Maintain a Data Warehouse

Here are some typical maintenance tasks for a data warehouse.


Exploit a Data Warehouse

Generally, SAS/Warehouse Administrator is used to build a data warehouse, not exploit it. Its main role is to create and manage warehouse data stores that are then accessed by other applications. In some cases, however, you can use SAS/Warehouse Administrator to prepare data or metadata for exploitation.

For example, here is one way to use SAS/EIS to analyze a summary data store.

  1. Use SAS/Warehouse Administrator to export metadata from the relevant summary data store to a SAS/EIS metabase. For details about this task, see Example: Exporting Metadata to SAS/EIS Software.

  2. In SAS/EIS, design the report you need.

  3. Run the report.

Another example is to surface a copy of the metadata for a given Data Warehouse. This method could be used to support business analysts who need to answer such questions as What are the subject areas in this Data Warehouse? What data stores are available for analysis? What is the origin of the data in this column? What does this column mean? How is it derived?

  1. Export metadata from the Data Warehouse. For details, see Exporting Metadata for Groups and Data Stores.

  2. Download the MetaSpace Explorer from the SAS Web site, as described in MetaSpace Explorer.

  3. Install and configure the MetaSpace Explorer to enable access to the exported metadata. For details about this task, see the documentation for the MetaSpace Explorer.

Previous Page | Next Page | Top of Page