Simple Input Validation through HTML Attributes

The best place to validate user input is at the source. If you can keep field validations in the browser, you can avoid unnecessary server requests and provide immediate user feedback while reducing the amount of bad data being submitted. This can be especially useful in the standard interactivity client because it can perform validations immediately, without waiting for the user to attempt to commit the record.

A Siebel form applet is an HTML form, and the controls are input elements on those forms. The HTML Attributes property of a Control object provides an opportunity to insert a JavaScript event to the input element.

For example, I recently implemented a query applet with a field validation on the Social Security Number control. To do so, I updated the HTML Attributes property of the control to

onkeyUp="if(/[^0-9\-]/.test(this.value) ){ alert('The Social Security Number field accepts only numeric data.'); this.value='';}"

The above validation intercepts the onkeyUp event of the input element, and uses a regular expression test to detect if the key pressed was any other besides a number or a hyphen. If an errant character is found, the applet displays a message and clears the control. This validation occurs entirely on the browser, but without adding any browser scripts to the applet.

Here's another example of a date validation on another control on the same applet:

onBlur="if(/^( *)((0[1-9]|[1-9]|1[012])[/](0[1-9]|[1-9]|[12][0-9]|3[01])[/](18|19|20)\d\d)( *)$/.test(this.value)||this.value==''){}else{alert('The Date of Birth field accepts date input in the M/D/YYYY or MM/DD/YYYY format');this.value='';}"

A somewhat more complex regular expression tests the format of a date. In this case, the expression doesn't test the input until the focus moves out of the field.

The HTML Attributes property of the Control object provides a clean, declarative approach for input validation.

A Regular Expression Validation

Data Validation Manager is great, but what would validation be without Regular Expressions? Here's how I used both to meet a business requirement.

The requirement arose from some errors logged when Siebel would attempt to send email to invalid addresses. We wanted to notify a user who entered invalid characters in the email address field. Of course, this edit wouldn't guarantee that addresses entered would actually exist, but it would at least ensure that people wouldn't separate two addresses with a colon (':') instead of a semicolon (';'), which is what was happening.

Step 1: Create a Business Service

Create a new Business Service with a method containing the following code snippet:

var sPattern = Inputs.GetProperty("Pattern");
var sSample = Inputs.GetProperty("Sample");
var rExp = new RegExp(sPattern);

if(rExp.test(sSample))
{
    Outputs.SetProperty("Result", 1);
}
else
{
    Outputs.SetProperty("Result", 0);
}

Step 2: Create a Data Validation Rule

Our Data Validation Rule Set is called "Employee Email Validation". In the data validation rule, use the new business service to test a business component field against a regular expression pattern. Use the following syntax for the pattern:

InvokeServiceMethod("ABC Validation Service", "Validate", "Pattern='^(([a-zA-Z0-9_\-\.]+)@([a-zA-Z0-9_\-\.]+)\.([a-zA-Z]{2,5}))+([;,](([a-zA-Z0-9_\-\.]+)@([a-zA-Z0-9_\-\.]+)\.([a-zA-Z]{2,5}))+)*$', Sample=eval([EMail Addr])", "Result")<>0

The regular expression pattern above was adapted from one available on the website http://www.regular-expressions.info/, where you can quickly learn how to get started with regular expressions. Also note the use of the eval() function, which will result in the contents of the Business Component Field being passed to the business service instead of the literal field name. See Metalink Doc ID 782338.1 (login required) for Oracle's How-To article for using eval() with InvokeServiceMethod() for a very similar application.

Step 3: Create a Runtime Event

Your next step is to create a Runtime Event manually through the Administration - Runtime Events screen. For our data validation, we created an Action Set for the Employee WriteRecord event.

• Action Type: BusService


• Business Service Name: Data Validation Manager


• Business Service Method: Validate


• Business Service Context: "Rule Set Name", "Employee Email Validation", "Enable Log", "Y"

Please note that although we did our validation on the Employee BusComp, it doesn't work on the Administration - User -> Employees. We have a custom view created on the Employee BusComp that exposes the email field. I'm not actually sure why it doesn't work in the vanilla view, but I'll post an update if I find out.

Step 4: Update Enterprise Parameter

A security feature introduced with Siebel 7.7 restricts the InvokeServiceMethod function to only business services that are registered on the Query Access List. The parameter can be set at the enterprise level or the component (e.g., Object Manager) level. However, since component-level settings always override enterprise settings, setting the parameter at the component level would preclude the use of any other business services that are registered at the enterprise.

• Enterprise Parameter Name: Business Service Query Access List


• Enterprise Parameter Value: ABC Validation Service

A common Siebel software limitation applies here. Only a maximum of 100 characters can be used to specify business service names in this parameter. You should carefully plan which business services you want to use with InvokeServiceMethod. Do not leave spaces between multiple business service names, but separate names with commas, and do not use quote marks.

Conclusion

After setting up a very simple business service to evaluate regular expressions, we can create any number of complex validations to display a message when a pattern is matched, with no additional scripting at all. For example, another requirement to validate the format of an identification number in a free text field was accomplished with a second data validation using the same business service.

If you are new to regular expressions, take a look. They are incredibly useful.