I love tuples. They allow you to quickly group relevant information together without having to write a struct or class for it. This is very useful while refactoring very localized code.
Initializing a list of them however seems a bit redundant.
var tupleList = new List<Tuple<int, string>>
{
Tuple.Create( 1, "cow" ),
Tuple.Create( 5, "chickens" ),
Tuple.Create( 1, "airplane" )
};
Isn't there a better way? I would love a solution along the lines of the Dictionary initializer.
Dictionary<int, string> students = new Dictionary<int, string>()
{
{ 111, "bleh" },
{ 112, "bloeh" },
{ 113, "blah" }
};
Can't we use a similar syntax?
This question is related to
c#
list
collections
.net-4.0
tuples
Old question, but this is what I typically do to make things a bit more readable:
Func<int, string, Tuple<int, string>> tc = Tuple.Create;
var tupleList = new List<Tuple<int, string>>
{
tc( 1, "cow" ),
tc( 5, "chickens" ),
tc( 1, "airplane" )
};
Yes! This is possible.
The { } syntax of the collection initializer works on any IEnumerable type which has an Add method with the correct amount of arguments. Without bothering how that works under the covers, that means you can simply extend from List<T>, add a custom Add method to initialize your T, and you are done!
public class TupleList<T1, T2> : List<Tuple<T1, T2>>
{
public void Add( T1 item, T2 item2 )
{
Add( new Tuple<T1, T2>( item, item2 ) );
}
}
This allows you to do the following:
var groceryList = new TupleList<int, string>
{
{ 1, "kiwi" },
{ 5, "apples" },
{ 3, "potatoes" },
{ 1, "tomato" }
};
One technique I think is a little easier and that hasn't been mentioned before here:
var asdf = new [] {
(Age: 1, Name: "cow"),
(Age: 2, Name: "bird")
}.ToList();
I think that's a little cleaner than:
var asdf = new List<Tuple<int, string>> {
(Age: 1, Name: "cow"),
(Age: 2, Name: "bird")
};
var colors = new[]
{
new { value = Color.White, name = "White" },
new { value = Color.Silver, name = "Silver" },
new { value = Color.Gray, name = "Gray" },
new { value = Color.Black, name = "Black" },
new { value = Color.Red, name = "Red" },
new { value = Color.Maroon, name = "Maroon" },
new { value = Color.Yellow, name = "Yellow" },
new { value = Color.Olive, name = "Olive" },
new { value = Color.Lime, name = "Lime" },
new { value = Color.Green, name = "Green" },
new { value = Color.Aqua, name = "Aqua" },
new { value = Color.Teal, name = "Teal" },
new { value = Color.Blue, name = "Blue" },
new { value = Color.Navy, name = "Navy" },
new { value = Color.Pink, name = "Pink" },
new { value = Color.Fuchsia, name = "Fuchsia" },
new { value = Color.Purple, name = "Purple" }
};
foreach (var color in colors)
{
stackLayout.Children.Add(
new Label
{
Text = color.name,
TextColor = color.value,
});
FontSize = Device.GetNamedSize(NamedSize.Large, typeof(Label))
}
this is a Tuple<Color, string>
You can do this by calling the constructor each time with is slightly better
var tupleList = new List<Tuple<int, string>>
{
new Tuple<int, string>(1, "cow" ),
new Tuple<int, string>( 5, "chickens" ),
new Tuple<int, string>( 1, "airplane" )
};
Super Duper Old I know but I would add my piece on using Linq and continuation lambdas on methods with using C# 7. I try to use named tuples as replacements for DTOs and anonymous projections when reused in a class. Yes for mocking and testing you still need classes but doing things inline and passing around in a class is nice to have this newer option IMHO. You can instantiate them from
- Direct Instantiation
var items = new List<(int Id, string Name)> { (1, "Me"), (2, "You")};
- Off of an existing collection, and now you can return well typed tuples similar to how anonymous projections used to be done.
public class Hold
{
public int Id { get; set; }
public string Name { get; set; }
}
//In some method or main console app:
var holds = new List<Hold> { new Hold { Id = 1, Name = "Me" }, new Hold { Id = 2, Name = "You" } };
var anonymousProjections = holds.Select(x => new { SomeNewId = x.Id, SomeNewName = x.Name });
var namedTuples = holds.Select(x => (TupleId: x.Id, TupleName: x.Name));
- Reuse the tuples later with grouping methods or use a method to construct them inline in other logic:
//Assuming holder class above making 'holds' object
public (int Id, string Name) ReturnNamedTuple(int id, string name) => (id, name);
public static List<(int Id, string Name)> ReturnNamedTuplesFromHolder(List<Hold> holds) => holds.Select(x => (x.Id, x.Name)).ToList();
public static void DoSomethingWithNamedTuplesInput(List<(int id, string name)> inputs) => inputs.ForEach(x => Console.WriteLine($"Doing work with {x.id} for {x.name}"));
var namedTuples2 = holds.Select(x => ReturnNamedTuple(x.Id, x.Name));
var namedTuples3 = ReturnNamedTuplesFromHolder(holds);
DoSomethingWithNamedTuplesInput(namedTuples.ToList());
C# 6 adds a new feature just for this: extension Add methods. This has always been possible for VB.net but is now available in C#.
Now you don't have to add Add()
methods to your classes directly, you can implement them as extension methods. When extending any enumerable type with an Add()
method, you'll be able to use it in collection initializer expressions. So you don't have to derive from lists explicitly anymore (as mentioned in another answer), you can simply extend it.
public static class TupleListExtensions
{
public static void Add<T1, T2>(this IList<Tuple<T1, T2>> list,
T1 item1, T2 item2)
{
list.Add(Tuple.Create(item1, item2));
}
public static void Add<T1, T2, T3>(this IList<Tuple<T1, T2, T3>> list,
T1 item1, T2 item2, T3 item3)
{
list.Add(Tuple.Create(item1, item2, item3));
}
// and so on...
}
This will allow you to do this on any class that implements IList<>
:
var numbers = new List<Tuple<int, string>>
{
{ 1, "one" },
{ 2, "two" },
{ 3, "three" },
{ 4, "four" },
{ 5, "five" },
};
var points = new ObservableCollection<Tuple<double, double, double>>
{
{ 0, 0, 0 },
{ 1, 2, 3 },
{ -4, -2, 42 },
};
Of course you're not restricted to extending collections of tuples, it can be for collections of any specific type you want the special syntax for.
public static class BigIntegerListExtensions
{
public static void Add(this IList<BigInteger> list,
params byte[] value)
{
list.Add(new BigInteger(value));
}
public static void Add(this IList<BigInteger> list,
string value)
{
list.Add(BigInteger.Parse(value));
}
}
var bigNumbers = new List<BigInteger>
{
new BigInteger(1), // constructor BigInteger(int)
2222222222L, // implicit operator BigInteger(long)
3333333333UL, // implicit operator BigInteger(ulong)
{ 4, 4, 4, 4, 4, 4, 4, 4 }, // extension Add(byte[])
"55555555555555555555555555555555555555", // extension Add(string)
};
C# 7 will be adding in support for tuples built into the language, though they will be of a different type (System.ValueTuple
instead). So to it would be good to add overloads for value tuples so you have the option to use them as well. Unfortunately, there are no implicit conversions defined between the two.
public static class ValueTupleListExtensions
{
public static void Add<T1, T2>(this IList<Tuple<T1, T2>> list,
ValueTuple<T1, T2> item) => list.Add(item.ToTuple());
}
This way the list initialization will look even nicer.
var points = new List<Tuple<int, int, int>>
{
(0, 0, 0),
(1, 2, 3),
(-1, 12, -73),
};
But instead of going through all this trouble, it might just be better to switch to using ValueTuple
exclusively.
var points = new List<(int, int, int)>
{
(0, 0, 0),
(1, 2, 3),
(-1, 12, -73),
};
Why do like tuples? It's like anonymous types: no names. Can not understand structure of data.
I like classic classes
class FoodItem
{
public int Position { get; set; }
public string Name { get; set; }
}
List<FoodItem> list = new List<FoodItem>
{
new FoodItem { Position = 1, Name = "apple" },
new FoodItem { Position = 2, Name = "kiwi" }
};
Source: Stackoverflow.com