Expression-Bodied Members Study Portal

A: A member whose body is a single expression after =>, instead of a statement block.

public class Person
{
    private string _name = "";

    public Person(string name) => Name = name;                        // constructor (C# 7)
    ~Person() => Console.WriteLine("finalized");                      // finalizer (C# 7)

    public string Name                                                // accessors (C# 7)
    {
        get => _name;
        set => _name = value ?? throw new ArgumentNullException(nameof(value));
    }

    public string Initials => $"{Name[0]}.";                          // read-only property (C# 6)
    public override string ToString() => $"Person {Name}";            // method (C# 6)
    public char this[int i] => Name[i];                               // indexer (C# 7)
    public static Person operator +(Person a, Person b)               // operator (C# 6)
        => new($"{a.Name}-{b.Name}");
}

Local functions and lambdas use the same form (int Twice(int x) => x * 2;).

A: C# 6: methods, read-only properties, indexers with only a getter, and operators (including conversion operators). C# 7.0: constructors, finalizers, get/set accessors (and later init, C# 9), and full indexers. Events' add/remove accessors can also be expression-bodied.

A: Yes, if the expression is a valid statement on its own: a method call, assignment, increment, await, or object creation. public void Log(string m) => _logger.LogInformation(m); is fine; public void X() => 42; is not.

A: No. public int Double(int x) => x 2; and public int Double(int x) { return x 2; } produce identical IL. The choice is purely about readability.

A: Only a get_X method; no backing field. The expression is evaluated every time the property is read. Compare the three forms:

public DateTime A { get; } = DateTime.Now;  // auto-property + initializer: stored once, at construction
public DateTime B => DateTime.Now;          // expression-bodied: recomputed on every read
public DateTime C = DateTime.Now;           // public FIELD with initializer (no property at all)

A returns the same time forever, B a new time on each read, and C is a field (no binary-compatible contract, not serialized by default). One character (= vs =>) changes the semantics, which makes this a favourite interview question.

A: Only when the expression is expensive or allocates. public decimal Total => _lines.Sum(l => l.Amount); walks the list every time; in a loop that reads Total repeatedly, cache it in a local, or store and update the total when the lines change. public IReadOnlyList<Line> Lines => _lines.AsReadOnly(); allocates a wrapper on every read; returning _lines typed as IReadOnlyList<Line> does not.

A:

A: When the body has several steps, side effects that deserve their own lines, error handling, or logging; when you would have to cram logic into nested ternaries to fit one expression; and when the line becomes hard to read. The team style rule (.editorconfig csharp_style_expression_bodied_methods etc.) usually says "when on a single line".

A: The result is usually the same, but the behaviour differs in edge cases. The version without async ("eliding" async/await) returns the inner task directly: it is slightly cheaper (no state machine), but:

Eliding is fine for pure one-line forwarding with nothing around it. As soon as there is a using, try or more than one call, use async/await.

A: The first is a read-only property that returns 5 (a method with no storage); the second is a public mutable field initialized to 5.

A: The shorthand public int X => …; is get-only. For a setter, write the accessors separately, each with its own expression body: public int X { get => _x; set => _x = Math.Max(0, value); }.

A: Yes. Since C# 7.0, throw is an expression in these positions: public void NotSupported() => throw new NotSupportedException(); and => value ?? throw ….

A: The caller adds to a brand-new list that is immediately discarded, because every read creates another list. Use { get; } = new(); for a stored list.