Revision as of 17:34, 29 November 2011

Access a database with EGL Rich UI

Lesson 6: Add code for the service functions

In EGL, I/O statements such as add and get access data that resides in different kinds of persistent data storage, from file systems to queues to databases. The coding is similar for the different cases.

In this lesson, you add functions that access rows in a relational database. Add the functions in order, before the final end statement in SQLService.egl.

Create binding to database connection

In Lesson 3, you defined a database connection named Derby'. Use these steps to use the Derby connection from your SQL service.

1. EGL projects have associated deployment descriptor (.egldd) files. Click on PaymentService in the project explorer and select Properties > EGLDevelopmentDeploymentDescriptor" to see that file PaymentService.egldd was created when you defined project Payment Service. Click on OK to close the properties window.
2. Open file PaymentService.egldd with the EGL employment descriptor editor.
3. Select the Resource Bindings folder
4. Add a new resource binding named DerbyResource of binding type SQL database binding.
5. Select radio button Reference the selected workspace connection below [retrieved at runtime].
6. Select Derby connection details.
7. Clink on Finish.
8. Close file PaymentServices.egldd

Using an SQL resource binding in service program

Insert a dataSource variable in the service program following the program name:

  package services;
  
  service SQLService
  	ds dataSource? { @Resource { bindingKey="Derby" } } ;   
  end

The syntax directs the service to use the binding named "Derby" defined as a resource in the deployment descriptor file associated with the service project.

Handling SQL exceptions

SQL operations can fail for a multitude of reasons. For our simple example, our service will catch all SQL and log all SQL exceptions on the server and then throw the exception back to the service client. It will also log each service invocation.

Cut and paste these logging function to the service program before the final end statement:

    logActive boolean = true;
    activeService string;
    
    private function logEntry(serviceFunction string in)
        activeService = serviceFunction;
        log("Entry:  SQLService, " + serviceFunction);
    end
    
    private function logException(ex sqlException?)
        accumulatedMessage string = "Exception:  SQLService, " + activeService;
        while(ex != null)
            accumulatedMessage = accumulatedMessage + ", SQLSTATE = " +
                    ex.SQLState + ", message = " + ex.message;
            ex = ex.nextException;
        end
        log(accumulatedMessage);
        throw new anyException{message = accumulatedMessage};
    end
    
    private function log(text string in)
        if(logActive)
            sysLib.writeStdOut(text);
        end
    end

Add a payment record

Include a function to the service program that uses the add statement to insert a new row to the database.

To code the function:

• In the EGL editor, copy and paste the following lines into SQLService.egl before

the end statement:

  function addPayment(newPayment paymentRec in)
       logEntry ( "addPayment" ) ;
       try
           add newPayment to ds ;
       onException(ex sqlException)
           logException(ex);
       end
  end

• Before you continue, you must resolve the reference to the paymentRec Record part. You can automatically create import statements by using the Organize Imports feature. Right-click any blank area in the editor and click EGL Source > Organize Imports.

EGL adds the following statement to the beginning of the file:

  import records.paymentRec;

The reference is now resolved. You will use this feature often, whether by selecting the menu item or by pressing Ctrl-Shift-O.

• Save the file (Ctrl-S), and then place your cursor anywhere in the add statement. Press keys CTRL-1 and select Add SQL Statement from the popup menu. This feature changes the implicit SQL that underlies the EGL add statement into embedded code that you can modify:

  function addPayment(newPayment paymentRec in)
       logEntry ( "addPayment" ) ;
       try
           add newPayment to ds
               with #sql{
                       insert into PAYMENT
                             (CATEGORY, DESCRIPTION, AMOUNT, FIXED_PAYMENT, DUE_DATE, 
                             PAYEE_NAME, PAYEE_ADDRESS1, PAYEE_ADDRESS2)
                       values
                             (?, ?, ?, ?, ?, ?, ?, ?)
                 };
       onException(ex sqlException)
           logException(ex);
       end
  end

Notice the PAYMENT_ID field is not included in the record fields added to the table. The @GeneratedValue annotation for the field in the record definition tells the EGL SQL builder not to explicitly add the column.

Related reference

[../../com.ibm.egl.lr.doc/topics/regl_data_sql_add.html add considerations for SQL] [../../com.ibm.egl.lr.doc/topics/regl_core_function.html Functions] [../../com.ibm.egl.lr.doc/topics/regl_core_import.html import] [../../com.ibm.egl.lr.doc/topics/regl_data_sql_overview.html SQL data access]

Read all database records

The getAllPayments function uses the get statement with an array object to read all of the records from the table and stores them in an array.

To code the function:

• In the EGL editor, copy and paste the following lines into SQLService.egl before

the end statement:

  function getAllPayments() returns(paymentRec[])
      payments paymentRec[];
      logEntry("getAllPayments");
      try
          get payments from ds;
      onException(ex sqlException)
          logException(ex);
      end
      return(payments);
  end

The EGL get statement generates an SQL SELECT statement to retrieve a result set. When the target of the get statement is a dynamic array of records, EGL retrieves all matching rows from the result set and inserts each successive row into the next array element.

• As with the add statement, you may place the cursor anywhere in the statement and press CNTL-1 to see the SQL statement EGL will use to retrieve the rows. You may again save the default statement and modify it if it does do what you want.
• Save the file.

Related reference

[../../com.ibm.egl.lr.doc/topics/regl_data_sql_get.html get considerations for SQL]

Replace a record

The editPayment function replaces an existing row in the database with an edited version. The default statement replaces all records in the table that have key values equal to the contents of the key fields in the record variable. Key fields are those fields declared with the @Id attribute (field payment_id in our example).

To code the function:

• In the EGL editor, copy and paste the following lines into SQLService.egl before the logging functions:

  function editPayment(chgPayment paymentRec in)
      logEntry("editPayment");
      try
          replace chgPayment to ds;
      onException(ex SQLException)
          logException(ex);
      end
  end

The EGL replace statement generates an SQL UPDATE statement.

• Save the file.

Related reference

[../../com.ibm.egl.lr.doc/topics/regl_data_sql_replace.html replace considerations for SQL]

Delete a record

The deletePayment function deletes the specified record from the table.

To code the function:

• In the EGL editor, copy and paste the following lines into SQLService.egl before the end statement:

  function deletePayment(delPayment paymentRec in)
      logEntry("deletePayment");
      try
          delete delPayment from ds;
      onException(ex SQLException)
          if(ex.SQLState != "02000") // sql state is five digits
              logException sqlException;
          end
      end
  end

• The EGL delete statement generates an SQL DELETE statement. If no rows are present, the Derby database returns an SQLState value of "02000", and the EGL runtime code throws an exception that the function catches: that is, processes,

in some onException logic.

• When a function catches but ignores an exception, processing continues without interruption. That rule applies to the preceding logic, when the value of SQLState is "02000". When a function uses the throw statement to throw an exception, the exception stays active. That rule also applies to the preceding logic, when the value of SQLState is other than "02000".

• At run time, if a service does not handle an exception, the service requester receives an exception of type ServiceInvocationException. Incidentally, if the service cannot be

accessed, the requester receives an exception of type ServiceInvocationException or ServiceBindingException, depending on the details of the error.

• Save the file.

Related reference

[../../com.ibm.egl.lr.doc/topics/regl_data_sql_delete.html delete considerations for SQL] [../../com.ibm.egl.lr.doc/topics/regl_core_xcpt.html Exception handling]

Create test data

The createDefaultTable function creates a set of data for testing your completed application.

To code the function:

• In the EGL editor, copy and paste the following lines into SQLService.egl before

the end statement:

  function createDefaultTable() returns(paymentRec[])
      payments paymentRec[];
      logEntry("createDefaultTable");
      try
          try
              execute from ds
                 with #sql{
                    delete from PAYMENT
             };
        onException(ex SQLException)
            if(ex.SQLState != "02000")  // sqlState is five digits
                throw ex;
            end
        end 
        ispDate date = dateTimeLib.dateFromGregorian(20140405);
        addPayment(new paymentRec{category = 1, description = "Apartment"
              , amount = 880, fixedPayment = yes, dueDate = new date
              , payeeName = "A Jones", payeeAddress1 = "100 Jones Dr"
              , payeeAddress2 = "Jonesboro, NC"
              });
        addPayment(new paymentRec{category = 2, description = "Groceries"
              , amount = 450, fixedPayment = no, dueDate = new date
              , payeeName = "B Jones", payeeAddress1 = "200 Jones Dr"
              , payeeAddress2 = "Jonesboro, NC"
          });
        addPayment(new paymentRec{category = 5, description = "ISP"
              , amount = 19.99, fixedPayment = no, dueDate = ispDate
              , payeeName = "C Jones", payeeAddress1 = "300 Jones Dr"
              , payeeAddress2 = "Jonesboro, NC"
        });
        payments = getAllPayments();
          onException(ex anyException)
              logException(ex);
          end
        return(payments);
    end

The code acts as follows:

• The EGL execute statement runs a literal SQL statement that deletes all rows from the PAYMENT table.
• The ispDate variable receives a date value from the dateTimeLib.dateValueFromGregorian() system function. The content of the variable is then in a format that is appropriate for insertion into the dueDate field in the database.
• The addPayment function is repeatedly invoked to add new rows to the PAYMENT table.
• The call to the getAllPayments function returns an array of rows that were retrieved from the table.

• Press Ctrl-Shift-F to format the code. If you see any red Xs, compare your code with the finished code in Code for PaymentFileMaintenance.egl after lesson 6
• Save and close the file.

Related reference

[../../com.ibm.egl.lr.doc/topics/regl_data_sql_execute.html execute considerations for SQL] [../../com.ibm.egl.lr.doc/topics/regl_core_date_date_value_greg.html dateValueFromGregorian()]

Lesson checkpoint

You learned how to complete the following tasks:

• Add embedded SQL code to a program and modify that code
• Automatically create and organize import statements

In the next lesson, you will create a widget to hold the table of expense data.