/// <summary>
/// Insert/update the customer
/// If it already exists, update it, otherwise insert it.
/// If the email address has changed, raise a EmailAddressChanged event on DomainEvents
/// </summary>
public void Upsert(Customer customer)
{
if (customer == null) { throw new ArgumentNullException("customer"); }
var db = new DbContext();
var existingDbCust = GetById(customer.Id);
var newDbCust = ToDbCustomer(customer);
if (existingDbCust == null)
{
// insert
db.Insert(newDbCust);
// Note that this code does not trap exceptions coming from the database. What would it do with them?
// Compare with the F# version, where errors are alway returned from the call
}
else
{
// update
db.Update(newDbCust);
// check for changed email
if (!customer.EmailAddress.Equals(existingDbCust.EmailAddress))
{
// Generate a event
// Note that this code is buried deep in a class and is not obvious
// It is also hard to turn on and off (eg for batch updates) without adding extra complications
//
// Compare this with the F# version, when an event is returned from the call itself.
DomainEvents.OnEmailAddressChanged(existingDbCust.EmailAddress, customer.EmailAddress);
}
}
}
/// <summary>
/// Return the customer with the given CustomerId, or null if not found
/// </summary>
public Customer GetById(CustomerId id)
{
var db = new DbContext();
// Note that this code does not trap exceptions coming from the database. What would it do with them?
// Compare with the F# version, where errors are returned along with the Customer
return db.Customers().Where(c => c.Id == id.Id).Select(FromDbCustomer).FirstOrDefault();
}
/// <summary>
/// Return all customers
/// </summary>
public IEnumerable<Customer> GetAll()
{
var db = new DbContext();
return db.Customers().Select(FromDbCustomer);
}