Exception handling

[andree surya]

Hi there,

I just want to confirm if I've done this the correct way.
I have a code to fetch a record from database, given the ID of the record.

/**
 * Get the transaction with the specified ID.
 */
public Transaction getById(int id)
	throws DataSourceException, DataNotFoundException
{	
	try
	{
		getByIdStatement.setInt(1, id);
		ResultSet result = getByIdStatement.executeQuery();
		if (! result.next())
			throw new DataNotFoundException();
		
		Transaction transaction = new Transaction();
		transaction.setId(result.getInt("id"));
		transaction.setDescription(result.getString("description"));
		transaction.setAmount(result.getDouble("amount"));
		transaction.setDate(result.getString("date"));
		
		result.close();
		return transaction;
		
	}
	catch (SQLException e)
	{
		System.err.println(e);
		throw new DataSourceException();
	}
	finally
	{
		getByIdStatement.clearParameters();
	}
}

The problem is, apparently the clearParameters() call in the finally block throws an SQLException.
I managed to solve this by putting another try catch clause inside the finally block.

try
{
	getByIdStatement.clearParameters();
}
catch (SQLException e)
{
	System.err.println(e);
	throw new DataSourceException();
}

Problem solved, by I hate this. Exception handling is cluttering up my code !
Can someone give an advice, how should I solve this problem? Or this is how it is, and I should just leave it be?

[Ninad Kulkarni]

@ andree
Suppose if we handled SQLException in clearParameters() method is that good idea to handle it there?
Let see what other will say about it.

[andree surya]

Ninad Kulkarni wrote:@ andree
Suppose if we handled SQLException in clearParameters() method is that good idea to handle it there?
Let see what other will say about it.

Thanks for the response Ninad.
I forgot to mention that getByIdStatement is a PreparedStatement object, included in the Java standard library.
I need to call clearParameters() method to the the object because it will be reused later.

[Mike Simmons]

Sometimes excessive-looking exception handling is just something you have to do in Java. In this example, it's possible to avoid the repeated catch SQLException simply by putting a single catch SQLException after everything else, which simplifies a bit. However I would also add another try/finally to close the ResultSet, just because it's good practice to close anything with a close() method, in a finally block, as soon as you're done with it:

/** 
 * Get the transaction with the specified ID. 
 */  
public Transaction getById(int id)  
    throws DataSourceException, DataNotFoundException {     
    try {  
        try {
            getByIdStatement.setInt(1, id);  
            ResultSet result = getByIdStatement.executeQuery();  
            try {
                if (! result.next())  
                    throw new DataNotFoundException();  
          
                Transaction transaction = new Transaction();  
                transaction.setId(result.getInt("id"));  
                transaction.setDescription(result.getString("description"));  
                transaction.setAmount(result.getDouble("amount"));  
                transaction.setDate(result.getString("date"));  
                return transaction;  
            }
            finally { result.close(); }
        }
        finally { getByIdStatement.clearParameters(); } 
    }  
    catch (SQLException e) { throw new DataSourceException(e); }   
}  

Note that I also didn't bother to print the SQLException, but I did include that exception in the DataSourceException constructor, which I think is more important. Whatever code catches that exception, it can print the full stack trace, and that will include the stack trace of the initial SQLException. That's useful information that we don't want to lose, and it's usually easier if, when the exception is eventually logged, all the stack trace info appears in one place.

[andree surya]

Ok. Thanks for your code and advice, Mike !