multi map

gistfile1.cs

cnn.QueryMultiple<Table<Order>,Table<OrderLine>,Order>(
@"select * from Orders
select * from OrderLine", (orders, lines) => 
{
   var map = lines.ToDictionary(line => line.OrderId);
   foreach(var order in orders)
   {
      order.Lines = map[o.Id];
      yield return order;
   }
});


// another option, infer Table for single: 

cnn.QueryMultiple<Table<Order,User>,OrderLine,Order>(
@"select * from Orders 
left join Users u on u.Id = OwnerId
select * from OrderLine", (orders, lines) => 
{
   var map = lines.ToDictionary(line => line.OrderId);
   foreach(var order in orders.Map((o,u) => {o.Owner = u; return o;}))
   {
      order.Lines = map[o.Id];
      yield return order;
   }
});


// Same code today 

grid = cnn.QueryMultiple(
@"select * from Orders 
left join Users u on u.Id = OwnerId
select * from OrderLine"); 

var orders = grid.Read<Order,User,Order>((o,u) => {o.Owner = u; return o;}).ToList();
var map = grid.Read<OrderLine>().ToDictionary(line => line.OrderId);

foreach(var order in orders)
{
  order.Lines = map[order.Id];
}

I went back and pulled up my existing code and compared it to this and my brain locked onto what you were asking and my jaw dropped. I'm not entirely sure I would appreciate this addition to the code, as I use it fairly differently. Here's one of my random methods, slightly anonymized:

    public static RecordDTO SprocName(string param1, string param2) {
        List<ThingIds> Things; /* the field returned is called ItemID */
        List<UnitSet> UnitSet;
        List<TemperatureData> Temps;
        List<UnitIntervals> UnitIntervals;

        using ( var con = DataAccessMaster.GetOpenConnection() ) {
            using ( var multi = con.QueryMultiple( "SprocName", new { Param1 = param1, Param2 = param2} ) ) {
                Things = multi.Read<ThingIds>().ToList();
                UnitSet = multi.Read<UnitSet>( ).ToList( );
                Temps = multi.Read<TemperatureData>( ).ToList( );
                UnitIntervals = multi.Read<UnitIntervals>( ).ToList( );
            }
        }
        return new RecordDTO {
            Things = Things,
            UnitSet = UnitSet,
            Temps = Temps,
            UnitIntervals = UnitIntervals
        };
    }

Isn't there a bug in the second option? Shouldn't it return o instead of u?

// another option, infer Table for single:

cnn.QueryMultiple<Table<Order,User>,OrderLine,Order>(
@"select * from Orders
left join Users u on u.Id = OwnerId
select * from OrderLine", (orders, lines) =>
{
   var map = lines.ToDictionary(line => line.OrderId);
   foreach(var order in orders.Map((o,u) => {o.Owner = u; return o;}))
   {
      order.Lines = map[o.Id];
      yield return order;
   }
});

If so that applies to the third example line as well

@william true ... changed it

@SamSaffron: .ToDictionary probably isn't right in any of the examples (why would we need multimap at all? :))

Can we get rid of even more boilerplate? For the multimap, I'd think all you really need to specify are the key selectors and the way to combine the results...

var results = cnn.Query<User, Order, OrderLine>
            ("select * from Users;select * from Order;select * from OrderLine")
            .Map((user, order) => { order.User = user; return order; })
            .MultiMap(order => order.Id, 
                      line => line.OrderId,
                      (order, orderLines) => order.OrderLines = orderLines.ToList());

@JasonPunyon isn't he trying to avoid chaining in his revision?

@jcolebrand: He hasn't told me so I wouldn't know. I just followed a tweet here :)

@JasonPunyon me too! ;)