Graphical Customer Order Query
Figure 1 shows how the BEA Liquid Data for WebLogic 8.1 query builder tool, Data View Builder (DVB), can be used to graphically construct the query that assembles all the information needed for Avitek's customer self-service information page. The shape of the query result is determined by its target schema (shown on the right); the target schema contains a top-level element called CustProfile. CustProfile has a child element, Customer, that has basic profile information such as Customer_ID, First_Name, Last_Name, Email_Address. The Customer element itself has a child element Customer_Order that contains summary information for any/all orders for a given customer. The customer profile information for the query comes from the Customer table in the relational data source RTL-Customer. The order information for a given customer results from calling the Web service method ElectOrderService.getOrderSummary(). This Web service call takes custid as a single input parameter that comes from the Customer_ID column of the Customer table. The query is designed to be a parameterized query that takes a customer id as input and returns the profile and order information for the indicated customer.

Customer Order XQuery Source
Listing 1 shows the detailed structure of the XQuery generated by Liquid Data for WebLogic 8.1 as a result of Figure 1's mappings. The topmost FOR clause in the XQuery binds a variable to each Avitek customer who satisfies the query's outer WHERE predicate, which only keeps those customers whose customer ID matches the given query parameter. The outermost RETURN clause specifies the query's output, which consists of customer data from the RTL-Customer database plus the result of a nested query. The nested query pulls order data from the Electronics Order Web Service. The nested query's WHERE clause ensures that only orders for the customer of interest are queried, and the inner RETURN clause retrieves the desired order information. Listing 2 shows a sample of this query's output.

Now that we have designed and tested the query successfully, we are ready to deploy the query and its associated target schema to the Liquid Data server as a stored query. Deploying a stored query directly from DVB is a new capability in Liquid Data 8.1. Once deployed, the query becomes immediately available for use by applications using the Liquid Data control in the WebLogic Workshop IDE, so we are now ready to create a Liquid Data Java control.

Creating a Liquid Data Control
Having deployed the Avitek customer order query using DVB, we are now ready to launch the BEA WebLogic Workshop IDE and start creating the desired application using the Liquid Data for WebLogic 8.1 control. To create an instance of a Liquid Data control, we will begin by creating a new default application in WebLogic Workshop and creating a folder called orders within it. We will then create a new control by selecting Liquid Data from the controls selection presented to us when we ask WebLogic Workshop to create a new Java control. Having done this, clicking on the next button will take us to a second screen that asks us to choose between a local and a remote Liquid Data server. Liquid Data supports two deployment scenarios. It can either be colocated in the same server as the final application, or it can be on a remote. In this example we will select the local option and click Create. This will take us to a third screen in the Liquid Data control creation process, which allows us to browse the deployed stored queries in the repository of the Liquid Data server. At this point we can select one or more queries to become methods when we create the Liquid Data control. Figure 2 shows the dialog for choosing among a set of available Liquid Data queries.

We can now select the query crm.CustOrder and press the Finish button; the Java source code for defining the desired Liquid Data control instance will be created as the file CustOrder.jcx. As alluded to earlier, creating this control also auto-generates the XMLBeans classes related to the target schema associated with the query. These XMLBeans classes provide typed accessor methods for traversing the XML data returned by the query. Listing 3 shows the contents of the Liquid Data control JCX file. Given that we selected just our one query when creating the control, this file defines just one method, one that corresponds to the query that we selected. The name of the method defaults to the name of the query, although developers have the option of changing the method name. Note that this method, CustomerOrder(), takes one parameter, Customer ID (the input parameter for the stored query CustomerOrder), and returns a result of type CustOrderDocument (the root document type for the collection of XMLBeans classes generated from the query's target XML schema). The method annotations in the JCX file include the name of the stored query and a copy of its XML schema.

Once this Liquid Data Java control instance has been created, it can be used in developing WebLogic Workshop applications such as page flows, portals, Web services, and workflows.

Creating Web Applications
BEA WebLogic Workshop 8.1 provides tools for developing Web applications using JavaServer Pages (JSPs) and page flows, separating presentation, business logic, and navigational control logic to help minimize application complexity. Page flows in WebLogic Workshop allow developers to separate their user interface code for a Web application from their navigational control logic and their other business logic (which may include data aggregation code). User interface code is placed where it belongs, in JSP files. Navigational control logic is implemented separately (but easily) in the page flow's controller file (which is a special Java file that uses a JPF file extension). Data aggregation logic can be implemented in a Liquid Data control. In this section we will walk through the process of creating a Web application that uses our Liquid Data control for data aggregation and page flows for data presentation.

Given a Liquid Data control, right-clicking on the control's JCX file causes WebLogic Workshop to prompt for one of two options: generate page flow or generate test JWS (Java Web service) file. If we select the option to generate a page flow, WebLogic Workshop will launch its page flow-generation wizard to guide us through the creation of a page flow based on our Liquid Data control. The default page flow that it generates includes JSPs that allow users to enter values for the input parameters for the control's methods and then to invoke the methods. The generated JSPs don't take care of result formatting, however; they simply display the returned results in XML string format. In this example, we want fancier formatting - we would like our result page to have a header to show the Customer profile data along with a table listing all the order information for the given customer.

To format the query's XML results as desired in the JSP result page, we will need to declare a public variable of type Customer in the page flow so that it will be accessible in the WebLogic Workshop page flow properties data palette. After declaring this variable and assigning the results of the query to it in the source code, we need to create a table for formatting header information. In the header block we'll simply create a table with two columns and four rows. The first column will have the labels for the Customer information and the second column will have the actual values. These values can be dragged and dropped into the table from the Customer object in the properties data palette. As shown in Figure 3, for the Order Information table we will need to drag-and-drop the CustomerOrderArray element from the Customer object in the properties data palette. As we drop the CustomerOrderArray onto the JSP, a RepeaterWizard will pop up and walk us through the creation of the orders table. The RepeaterWizard will provide us with choices, including the selection of what fields we want to display and the formatting style to be used. At this point we have completed the formatting of our JSP pages and are ready to test our application. Figure 4 shows the output of running the resulting Web application. It shows the results formatted with a header block containing the customer profile information and a detail block containing all of the orders related to the customer. It is interesting to note how little actual coding was required to create the desired customer self-service application.

Once we've created a Liquid Data control, it's just as easy to create a portal, a Web service, or a workflow that incorporates it. The paradigm for consuming Liquid Data queries in any other WebLogic Workshop project is the same as what we've shown for creating Web applications.

Other Notable Liquid Data Features
In addition to the Liquid Data control, which is the central new feature in Liquid Data for WebLogic 8.1, this release added several new ways for XML data to enter the Liquid Data world and improved on manageability and performance in several key areas.

Delimited Files
Liquid Data 1.0 supported only XML files when it came to file data sources. BEA Liquid Data for WebLogic 8.1 adds support for delimited files, which are text files that contain a predefined character as a data separator. Delimited files are typically exported from a database, a spreadsheet, or another application like a spreadsheet. It's now possible to create delimited file data sources in Liquid Data for use in queries. Such data sources have associated (partially implied) XML schemas and they make it possible to combine delimited file data with any other Liquid Data data source(s) by creating queries that use the delimited file data source.

Complex Parameter Types
Complex parameter types (CPTs) are another addition. Essentially, CPTs make it possible to use XML data as an input parameter to a Liquid Data query. CPTs make it possible to define an XML data stream of an arbitrary (XML Schema) type and then pass XML data of that type as input to a query. A typical CPT use case might be a workflow that receives an XML message and needs to enrich the message with information culled from a collection of back-end heterogeneous data sources.

SQL Calls as Data Sources
BEA Liquid Data for WebLogic 8.1 also includes support for using SQL stored procedures and SQL queries as parameterized data sources for use in Liquid Data queries and views. To use a stored procedure or a SQL query as a data source in Liquid Data, one creates a SQL Call Description File (SCDF). The SCDF is an XML file that defines the types and the functions for a set of SQL stored procedures and/or SQL queries. These SQL calls then become available for use in building queries in the Data View Builder just like any other data source.

Summary
In this article, we looked at how BEA Liquid Data for WebLogic 8.1 makes it easy to develop applications - including Web applications, portals, Web services, and workflows - that require access to data from multiple heterogeneous sources. You can try developing applications using BEA Liquid Data for WebLogic 8.1 yourself by downloading it from http://commerce.bea.com/index.jsp.

References