Menu Request Demo

Impact Analysis: What can fizzle while applying changes on the database and how to sustain a healthy, evolving data warehouse?

10 June 2008
Raphael Klebanov

Every action in the information database and warehouse in particular has a consequence. Database developers, business analysts and DBAs  feel agonizing pain when an organization makes changes to their data warehouse applications without having complete insight into the true impact and cost that a modification will have.

So, what is an impact analysis?  It is an identification of the consequences of change on an object to its related objects. Impact analysis, in case of a data warehouse, consists of:

• understanding of the processes performed within a data warehouse as a whole and its components;
• comprehension of interdependencies between processes and objects;
• assessing the effect the change brings to the data warehouse structure;
• developing recovery procedures in case of  a failed modification.

WhereScape RED provides a variety of needed mechanisms to identify a data warehouse’s weak spot when it comes to a change.

In the development process we have put our best effort into creating a perfect addition to our data warehouse. Now it’s time to deploy our changes to a higher environment. But how can we do so without crashing things that we had developed previously? How do we insure that the new modification will not destroy the integrity of the data warehouse instead of improving it? With WhereScape RED it’s easy.

Here are a few recommendations on how to avoid oopses and ouches:

1. Structural Impact Analysis.  i.e.  what objects have been changed.
Click “Tools” >  Click “Search for string…” > Fill out dialog box for string (e.g. either enter the procedure name being modified or the column being changed) > examine “Object search results” screen for all objects that contain this string.

For the structural-impact analysis you can use the Browser screen for "DataWarehouse" connection.
Click “Invoke SQL Admin” > execute something like this (SQL Server):

select distinct( from sys.procedures a inner join sys.syscomments b on
where charindex(‘stored procedure or column’,b.text)<>0;
-- Check load where clause too
select distinct(a.lt_table_name)from ws_load_tab a
where charindex(‘stored procedure or column’,a.lt_where_clause)<>0;

2. Evaluation of the Impact. i.e. what possible impact the data warehouse’s alteration renders on various data warehouse objects. WhereScape uses an array of pre-defined reports. Click the “Reports” tab from the main Browser menu. The list of available reports expands from release to release. There are about two dozen reports that are directly related to Impact Analysis. Some of them are:
•  “Object modification dates (without indexes)” with date created and date modified, same for “Index modification dates”;
• “Modified procedures” i.e. procedures that have been modified since creation;
• “Table objects that have no associated procedure or script” and “Procedures that are not related to a table”;
• “All objects that have been Refreshed or Imported”;
• “Differences between current and selected repository…”;
• “Query data warehouse objects… where custom validations scripts can be run”;
• “Track back on specified column usage…” and many others.

3. Impact during the Promotion from one environment to the other. e.g.  from DEV to TEST,  when managing such deployments using the “Build Application tables…” built-in utility. While the developer plunges through the wizard-like dialog boxes, WhereScape RED continuously warns him/her of the possible implications of this activity. E.g. if the particular object is to be overwritten, warning message pops up to alert the developer on the possible impact on the existing object.

4. Organizational Documentation Update. The considerable problem for a conventional data warehousing project is that database changes within data warehouse are not reflected in documentation: nobody has time to update them. WhereScape RED automatically generates a new documentation set after each iteration of data warehousing development.

5. Versioning of Metadata Objects. Several kinds of version management are utilized in WhereScape RED. For example, choosing the "Automated Version Creation" option from the Tools menu will thwart a developer from losing valuable DDL scripts, procedural code and other metadata and database information.

6. Standardization and Best Practice Usage.    WhereScape RED uses a state-of-the-art methodology in building warehousing prototypes. Following the developer’s guidelines delineated by the WhereScape RED methodology, wizards and the automatic generation of database objects, in combination with an enterprise standard “look and feel” dramatically decreases the chances of glitches in the data warehouse processing.


Comments 0

Leave a Comment

No comments.