[Proposal]: Collection expressions (VS 17.7, .NET 8)

[CyrusNajmabadi]

Collection expressions

• Proposed
• Specification: https://github.com/dotnet/csharplang/blob/main/proposals/csharp-12.0/collection-expressions.md
• Prototype: Complete
• Implementation: In progress. Shipping everything but 'natural types' and 'dictionary literals' in C# 12.

Many thanks to those who helped with this proposal. Esp. @jnm2!

Summary

Collection expressions introduce a new terse syntax, [e1, e2, e3, etc], to create common collection values. Inlining other collections into these values is possible using a spread operator .. like so: [e1, ..c2, e2, ..c2]. A [k1: v1, ..d1] form is also supported for creating dictionaries.

Several collection-like types can be created without requiring external BCL support. These types are:

• Array types, such as int[].
• Span<T> and ReadOnlySpan<T>.
• Types that support collection initializers, such as List<T> and Dictionary<TKey, TValue>.

Further support is present for collection-like types not covered under the above, such as ImmutableArray<T>, through a new API pattern that can be adopted directly on the type itself or through extension methods.

Motivation

• Collection-like values are hugely present in programming, algorithms, and especially in the C#/.NET ecosystem. Nearly all programs will utilize these values to store data and send or receive data from other components. Currently, almost all C# programs must use many different and unfortunately verbose approaches to create instances of such values. Some approaches also have performance drawbacks. Here are some common examples:

  • Arrays, which require either new Type[] or new[] before the { ... } values.

  • Spans, which may use stackalloc and other cumbersome constructs.

  • Collection initializers, which require syntax like new List<T> (lacking inference of a possibly verbose T) prior to their values, and which can cause multiple reallocations of memory because they use N .Add invocations without supplying an initial capacity.

  • Immutable collections, which require syntax like ImmutableArray.Create(...) to initialize the values, and which can cause intermediary allocations and data copying. More efficient construction forms (like ImmutableArray.CreateBuilder) are unweildy and still produce unavoidable garbage.

• Looking at the surrounding ecosystem, we also find examples everywhere of list creation being more convenient and pleasant to use. TypeScript, Dart, Swift, Elm, Python, and more opt for a succinct syntax for this purpose, with widespread usage, and to great effect. Cursory investigations have revealed no substantive problems arising in those ecosystems with having these literals built in.

• C# has also added list patterns in C# 10. This pattern allows matching and deconstruction of list-like values using a clean and intuitive syntax. However, unlike almost all other pattern constructs, this matching/deconstruction syntax lacks the corresponding construction syntax.

• Getting the best performance for constructing each collection type can be tricky. Simple solutions often waste both CPU and memory. Having a literal form allows for maximum flexibility from the compiler implementation to optimize the literal to produce at least as good a result as a user could provide, but with simple code. Very often the compiler will be able to do better, and the specification aims to allow the implementation large amounts of leeway in terms of implementation strategy to ensure this.

An inclusive solution is needed for C#. It should meet the vast majority of casse for customers in terms of the collection-like types and values they already have. It should also feel natural in the language and mirror the work done in pattern matching.

This leads to a natural conclusion that the syntax should be like [e1, e2, e3, e-etc] or [e1, ..c2, e2], which correspond to the pattern equivalents of [p1, p2, p3, p-etc] and [p1, ..p2, p3].

A form for dictionary-like collections is also supported where the elements of the literal are written as k: v like [k1: v1, ..d1]. A future pattern form that has a corresponding syntax (like x is [k1: var v1]) would be desirable.

Detailed design

The content of the proposal has moved to proposals/collection-expressions.md. Further updates to the proposal should be made there.

Design meetings

https://github.com/dotnet/csharplang/blob/main/meetings/2021/LDM-2021-11-01.md#collection-literals
https://github.com/dotnet/csharplang/blob/main/meetings/2022/LDM-2022-03-09.md#ambiguity-of--in-collection-expressions
https://github.com/dotnet/csharplang/blob/main/meetings/2022/LDM-2022-09-28.md#collection-literals
https://github.com/dotnet/csharplang/blob/main/meetings/2023/LDM-2023-04-03.md#collection-literals
https://github.com/dotnet/csharplang/blob/main/meetings/2023/LDM-2023-04-26.md#collection-literals
https://github.com/dotnet/csharplang/blob/main/meetings/2023/LDM-2023-05-03.md#collection-literal-natural-type
https://github.com/dotnet/csharplang/blob/main/meetings/2023/LDM-2023-05-31.md#collection-literals
https://github.com/dotnet/csharplang/blob/main/meetings/2023/LDM-2023-06-05.md#collection-literals
https://github.com/dotnet/csharplang/blob/main/meetings/2023/LDM-2023-06-19.md#collection-literals
https://github.com/dotnet/csharplang/blob/main/meetings/2023/LDM-2023-07-12.md#collection-literals
https://github.com/dotnet/csharplang/blob/main/meetings/2023/LDM-2023-09-18.md#collection-expression-questions
https://github.com/dotnet/csharplang/blob/main/meetings/2023/LDM-2023-09-20.md#collection-expressions
https://github.com/dotnet/csharplang/blob/main/meetings/2023/LDM-2023-09-25.md#defining-well-defined-behavior-for-collection-expression-types
https://github.com/dotnet/csharplang/blob/main/meetings/2023/LDM-2023-09-27.md#collection-expressions
https://github.com/dotnet/csharplang/blob/main/meetings/2023/LDM-2023-10-02.md#collection-expressions
https://github.com/dotnet/csharplang/blob/main/meetings/2023/LDM-2023-10-11.md#collection-expressions
https://github.com/dotnet/csharplang/blob/main/meetings/2023/LDM-2023-11-15.md#nullability-analysis-of-collection-expressions
https://github.com/dotnet/csharplang/blob/main/meetings/2024/LDM-2024-01-10.md
https://github.com/dotnet/csharplang/blob/main/meetings/2024/LDM-2024-02-26.md#collection-expressions

Working group meetings

https://github.com/dotnet/csharplang/blob/main/meetings/working-groups/collection-literals/CL-2022-10-06.md
https://github.com/dotnet/csharplang/blob/main/meetings/working-groups/collection-literals/CL-2022-10-14.md
https://github.com/dotnet/csharplang/blob/main/meetings/working-groups/collection-literals/CL-2022-10-21.md
https://github.com/dotnet/csharplang/blob/main/meetings/working-groups/collection-literals/CL-2023-04-05.md
https://github.com/dotnet/csharplang/blob/main/meetings/working-groups/collection-literals/CL-2023-04-28.md
https://github.com/dotnet/csharplang/blob/main/meetings/working-groups/collection-literals/CL-2023-05-26.md
https://github.com/dotnet/csharplang/blob/main/meetings/working-groups/collection-literals/CL-2023-06-12.md
https://github.com/dotnet/csharplang/blob/main/meetings/working-groups/collection-literals/CL-2023-06-26.md
https://github.com/dotnet/csharplang/blob/main/meetings/working-groups/collection-literals/CL-2023-08-03.md
https://github.com/dotnet/csharplang/blob/main/meetings/working-groups/collection-literals/CL-2023-08-10.md

[YairHalberstadt]

Overall looks good! Just a couple of points.

This facility thus prevents general use of such a marked method outside of known safe compiler scopes where the instance value being constructed cannot be observed until complete.

In the context of collection literals, the presence of these methods would allow types to trust that data passed into them cannot be mutated outside of them, and that they are being passed ownership of it. This would negate any need to copy data that would normally be assumed to be in an untrusted location.

This only works because at the moment it happens to be only a compiler can call init methods, if you don't use them yourself in one of your init properties.

However it doesn't seem like the sort of thing we'd want to rely on not changing in the future. For example we might allow calling init methods in the object initializer, at which point this would no longer be safe:

int[] ints = [1,2,3];
var immutArray = new{ Init(ints) };
int[0] = 5;

To resolve this, we could say that when evaluating a literal's spread_element expression, that there was an implicit target type equivalent to the target type of the literal itself. So, in the above, that would rewritten as:

It seems like this is not the most efficient solution - instead we would want to effectively inline the element, never materializing them into an array in the first place, and instead storing all the sub_elements on the stack.

Should we expand on collection initializers to look for the very common AddRange method? It could be used by the underlying constructed type to perform adding of spread elements potentially more efficiently. We might also want to look for things like .CopyTo as well. There may be drawbacks here as those methods might end up causing excess allocations/dispatches versus directly enumerating in the translated code.

We could only use those methods when the type of the spread_element exactly matches the parameter type of these methods, meaning we would not be causing virtual dispatch.

Can an unknown length literal create a collection type that needs a known length, like an array, span, or Init(array/span) collection? This would be harder to do efficiently, but it might be possible through clever use of pooled arrays and/or builders.

This is an extremely common use case - ToArray is the most commonly used Linq method. It would be unfortunate if the one syntax to rule them all required you to do: ((List<int>)[1, .. subspread, 2]).ToArray().

[CyrusNajmabadi]

@YairHalberstadt All good points. Thank you :)

[alrz]

Linking to https://github.com/bartdesmet/csharplang/.../proposals/params-builders.md on the builder pattern.

[orthoxerox]

What if collection literals with an unknown length have a different natural type from those with a known length? The latter should likely be a List<T>, but the latter is an IEnumerable<T> with an unspeakable implementation? That is, [..c1, ..c2] where either collection has an unknown length is semantically equivalent to Enumerable.Concat(c1, c2).

[jnm2]

@orthoxerox Once you're able to target-type literals of unknown length to T[] and ImmutableArray<T>, which I very much hope is made possible and which is not solved by giving such literals a different natural type, then will it still be advantageous to give them a different natural type?

[erikhermansson79]

A downside to using an array would be if a natural type is added for collection literals and that natural type is not T[]. There would be a potentially surprising difference when refactoring between var x = [1, 2, 3]; and IEnumerable x = [1, 2, 3];.

Can you explain this to me, please? Wouldn't whatever type you choose as a natural type implement IEnumerable<T>? What would the difference be?

[jnm2]

@erikhermansson79 Besides is type checks and pattern matching behaving differently, as well as GetType, and overload resolution if you use dynamic, there's also this for example: casting to IList and reading IsFixedSize will be observably different. If you place this in a public IEnumerable<int> property and databind using a UI framework, the difference in behavior could amount to a user-facing regression. And so on.

[orthoxerox]

@orthoxerox Once you're able to target-type literals of unknown length to T[] and ImmutableArray<T>, which I very much hope is made possible and which is not solved by giving such literals a different natural type, then will it still be advantageous to give them a different natural type?

I agree, enumerating the collection literal just by changing the type of the variable it's assigned to sounds confusing.

[MgSam]

I think proposals benefit from having a section of various examples. Most people can't look at grammar rules and get a good feel for what the syntax will actually look like.

Separately, I will slightly object to this statement,

Looking at the surrounding ecosystem, we also find examples everywhere of list creation being more convenient and pleasant to use. TypeScript...

TS/JS just has a different syntax for initializing arrays. I don't know that its really much more convenient than new[] { ... }. It certainly doesn't have a generalized collection initialization syntax whatsoever. If you use a collection type other than Array in JS you are SOL.

[CyrusNajmabadi]

I think proposals benefit from having a section of various examples

It will look like: [a, b, .. c, d]

[CyrusNajmabadi]

Separately, I will slightly object to this statement,

The statement was that the presence of this literal form has not proven to itself be problematic for these languages. Not that the literal is sufficient for all usages in those languages.

In other words, these literals are not in "the bad parts". Nor are there contingents if users recommending people not use these.

[jnm2]

Here's an immediate example I can think of:

var processArguments =
[
    "pack",
    "-o", outputPath,
    ..(configuration is not null ? (["-c", configuration]) : []),
    "/bl:" + Path.Join(artifactsDir, @"logs\pack.log"),
];

[CyrusNajmabadi]

I'm pretty firmly against approaches that either:

1. Involve lazy evaluation. For that, use linq.
2. Change the natural type of the expression based on the values within. I think that will just be too confusing and difficult to reason about.

[bernd5]

Why not use curly braces like we have them for arrays already?

So you could write:

List<int> myInts = {1, 2, 3, 4};

[CyrusNajmabadi]

Why not use curly braces like we have them for arrays already?

This is referred to in the motivation section, but i'll give a little more information. Effectively {s have issues for us in being ambiguous in some situations if they mean lists or properties. So we've moved to [ with list patterns to have no ambiguity and to give a nice and clean syntax for list-like things. When we designed that we were aware that we likely wanted a 'literal correspondence' as well, which this proposal is.

I also cover the move from { to [ in the drawbacks section, albeit with the idea that this would allow us to have uniformity everywhere moving foward.

ALso note that { as an expression-form is potentially highly destabalizing for future work. For example, if we ever want a block expression then we couldnt' have that use { if { is also for lists. We sketched out and very much liked [...] for list patterns and these literals (both for looks, and for sidestepping a lot of issues), so we made the decision back in the pattern design to go this direction. We simply ordered it as "list pattern in c#10" and now hopefully "list literal in c#11".

[TahirAhmadov]

Why not follow with your previous proposal though?

var sc = new SomeControl
{
  StringsOptions += ["b", "c", .. options]
};

Because it's not yet available :)

PS. I decided to table this for now. The problem is that if I add a Add(IEnumerable) overload (in any way - extension, sub-class, struct, doesn't matter), then I can't do new SomeControl { Items = { new("abc") } } anymore - the new() syntax fails to compile because IEnumerable<T> cannot be instantiated, obviously. Unfortunately, I did wholesale apply the fixer to use new() everywhere, so it would be too big of a breaking change for me.
However, others may find this useful. Still, I really hope the +=[ ... ] syntax (and the #5176) are added sooner than later.

[OJacot-Descombes]

Idea: flexible natural type:

var a = [1, 2, 3];     // int[]
var b = [1, 2, 3, ..]; // List<int>

[colejohnson66]

If you want a definite type, just use that type instead of var.

[OJacot-Descombes]

If you want a definite type, just use that type instead of var.

There are other situations, e.g., where you have to pass an argument to a parameter of type object or IList where you are not in control of the target type. If you want to use a definite type you fall of the cliff and would have to relay on a new expression with a collection initializer. My Idea is about being able to use a collection expression and still have some control of the outcome.

[mrwensveen]

If you want a definite type, just use that type instead of var.

There are other situations, e.g., where you have to pass an argument to a parameter of type object or IList where you are not in control of the target type. If you want to use a definite type you fall of the cliff and would have to relay on a new expression with a collection initializer. My Idea is about being able to use a collection expression and still have some control of the outcome.

I see your point, but List<int> b = [1, 2, 3]; already works, so it would be confusing. Would

List<int> b = [1, 2, 3];
List<int> b = [1, 2, 3, ..];

mean the same? Probably, yes?

I think the idea of a suffix has been floated already, and I can see it making sense in the scenario described above (where the target type can't be inferred). For example A for array and L for list (L already exists for number primitives, but I'd say the context is clear enough).

static void DoStuff(object myObj) => Console.WriteLine(myObj);
DoStuff([1, 2, 3]A); // "System.Int32[]"
DoStuff([1, 2, 3]L); // "System.Collections.Generic.List`1[System.Int32]"

[miyu]

Type suffixes for collection expressions seem interesting. If they were implemented I'd like to see them support tuples as well (or be overloaded by item count)... They're a natural extension to numeric literal suffixes (1.0f, 123.4m) and would be useful for creating types that feel like literals (e.g. a Vector3 (0, 0, 0)f == 0f3 vs Vector4 (0, 0, 0, 0)f == 0f4). Bonus points if they can be used with source generators for compile-time codegen. 100% people would then start using them as postfixed invocations though (e.g. [a, b, c]min).

I'd like to see them exposed via attributes as follows:

namespace System.Collections.Generic;

public class List<T> {
    // Option 1
    [Suffix("L")] 
    public List(...) {}

    // Option 2
    [Suffix("L")] 
    public static List<T> FromCollectionExpression(...) {}
}

public static class ListHelper { 
    // Option 3
    [Suffix("L")] 
    public static List<T> FromCollectionExpression(...) {}
}

Using the literal would require using System.Collections.Generic to pull the converter into scope.

As an alternative, the converter could be a class:

namespace System.Collections.Generics;

// Option 4
[Suffix("L")]
public static class ListSuffix {
    public static List<T> Convert(...) {}
}

And used the same way:

using System.Collections.Generic;

[1, 2, 3]L

With the benefit of supporting using aliases?:

using zz = System.Collections.Generic.ListSuffix;
[1, 2, 3]zz

The largest drawback, IMO, is that supporting empty-collections might still be ugly with this syntax & you potentially end up moving typing from the collection-creation to the item add in many cases... What's the empty hashset or empty dictionary? []hs<int> and []dict<string, int>?...

[TonyValenti]

Personally, I'd prefer to see special overloads of To{CollectionType} that can be used on collection expressions and have the same effect as a cast/target typing.

[mrwensveen]

Personally, I'd prefer to see special overloads of To{CollectionType} that can be used on collection expressions and have the same effect as a cast/target typing.

ToXX() implies a conversion and would clash with methods in System.Linq.Enumerable. AsXX() would be better, but how would you prevent existing extension methods from being called? This has the potential to break a lot of existing code.

[TahirAhmadov]

Regardless of the extension method name, I don't think there is an answer to the question of what to do with spreads, are they all pushed to the stack and a Span<T> is passed to the extension method? Or if we go with IEnumerable<T>, it creates an extra allocation. In both cases, now we have to enumerate the spreads once to populate the temporary buffer, and once more to create the actual collection. (However it may be possible to optimize it all away, not sure about that.)

The suffixes are a little too narrow-scoped.

I think the best solution to this would be generic arguments inference:

List<> list = [1,2,3];
ImmutableArray<> ia = [1,2,3];
Dictionary<,> dict = [1:"a", 2:"b", 3:"c"];

And it can be useful for other scenarios, too.

WRT var, I still think it should be either List<T> or Dictionary<K,V>, respectively, and not T[] or any other type.

[jnm2]

If you want to use a definite type you fall of the cliff and would have to relay on a new expression with a collection initializer.

@OJacot-Descombes Actually, it's not that drastic. You can always specify a target type for any expression, including collection expressions, using explicit cast syntax: (List<int>)[1, 2, 3].

[CreateAndInject]

Is natural types of Collection expressions is still developing?

[CyrusNajmabadi]

Yes

[CreateAndInject]

Yes

@CyrusNajmabadi Can you add to working set? https://github.com/dotnet/roslyn/blob/main/docs/Language%20Feature%20Status.md

[CyrusNajmabadi]

I don't think it's in the working set currently.

[mungojam]

You can already do ImmutableArray<int> xs = [1, 2, 3]; or var xs = (ImmutableArray<int>)[1, 2, 3];
I.e. target typing.

Thanks, I know and use target typing but the idea of natural typing, probably coming in .net 10 is allowing:

var x = [1, 2, 3];

But given that our team standardises on something other than List, we won't get to use this feature with the way it is currently proposed

[mrwensveen]

I see your point. You could probably write a source generator that allows you to write var x = [1, 2, 3]; and turns it into WhateverYourTeamLikes<T> x = [1, 2, 3];

[BreyerW]

@mrwensveen im not sure how you can achieve that specifically with current source generators. They are additive only by design and anything transformative like turning var into something concrete is nonstarter.

That aside i wonder if an attribute (assembly wide? or file wide? or any other scope?) that allows pointing to compiler which type should be treated as default for var collection expression would be palatable enough to cover situations like these?

[jnm2]

I see a lot of value in ImmutableArray for collections that outlive a method body, but not so much for other collections. You can see all the mutations (if any) right there in the method. It's not fundamentally different than a += on an int in terms of mutability of local values.

[mungojam]

It's not fundamentally different than a += on an int in terms of mutability of local values.

Something we also pretty much outlaw in our pseudo functional programming practices. And it looks like the idea of a local const or read-only declaration is being discussed again to help

[CreateAndInject]

Can natural types finish within 2025?

[huoyaoyuan]

It hasn't started. The focus of this version contains dictionary expression.

[333fred]

And collection expression arguments (which is important most specifically to dictionaries, but will be applicable to all collection expressions).