Bookshelf Home | Contents | Index | Search | PDF

Configuration Guidelines > Performance Guidelines >

Generating and Analyzing SQL

---

After configuring a view, you should always check the SQL that is generated during runtime. Use the /s filename parameter on the command line, as shown in the following example, to start the Siebel eBusiness application. This spools the generated SQL to a file. If you do not specify a path, the file is created in the Siebel root bin directory c:\siebel\bin. This file has all of the unique SQL statements generated during the current session and is overwritten during every new session.

The following example shows a command line using the /s filename parameter:

D:\siebel\core6\BIN\siebel.exe /c "D:\siebel\core6\bin\uagent.cfg" /s "c:\siebel.doc"

NOTE:  The spool file is simply a text file holding spooled SQL. Spooling it into a DOC file correctly formats it for viewing.

Once generated, analyze the contents of the file to identify any possible performance issues. Key indicators are:

Complexity of SQL statements
Order By Clause
Where Clause
Joins

Next, execute potentially problematic queries directly against the database and then generate a Query Plan. Use this plan to determine whether:

Indexes are being used
Tables Scans are occurring
Temporary Tables are being generated

Finally, comparing with a standard application lets you identify any potentially slow queries.

You can resolve many performance issues either by modifying the business component or an applet's search specification and the business component's sort specification, or by creating new indexes on the base table. Only specially trained Siebel Systems personnel can modify existing Siebel indexes. This restriction is enforced so that performance in other modules (such as the Enterprise Integration Manager) is not adversely affected by any index modifications you make to improve query performance through the user interface.

Consider any potential performance implications before modifying the search specification and sort specification properties for a business component. By spooling out the SQL, as previously described, you can analyze which indexes are likely to be used when your application queries the business component, through each applet.

Also, use your Relational Database Management System (RDBMS) vendor's tools to analyze the SQL your implementation generated.

Bookshelf Home | Contents | Index | Search | PDF