Install the package via NuGet:
dotnet add package FastCloner
dotnet add package FastCloner.Contrib # only required for some special types, such as Fonts
Clone your objects:
using FastCloner.Code;
var clone = FastCloner.DeepClone(new { Hello = "world", MyList = new List<int> { 1 } });
⭐ That's it! Feel free to map this method to your extension so if you need to migrate in the future it's a matter of just switching that method. We intentionally don't ship our own .DeepClone()
extension method.
Sometimes you might want to exclude certain fields & properties from cloning:
private class TestPropsWithIgnored
{
[FastClonerIgnore] // <-- decorate such members with [FastClonerIgnore]
public string B { get; set; } = "My string";
public int A { get; set; } = 10;
}
TestPropsWithIgnored original = new TestPropsWithIgnored { A = 42, B = "Test value" };
TestPropsWithIgnored clone = original.DeepClone(); // clone.B is null (default value of a given type)
Apart from deep cloning, FastCloner supports shallow cloning and deep cloning to target:
// the list is shared between the two instances
var clone = FastCloner.ShallowClone(new { Hello = "world", MyList = new List<int> { 1 } });
FastCloner uses caching by default which makes evaluating properties harder. Cloning unmanaged resources, such as IntPtr
s may result in side-effects, as there is no metadata for the length of buffers such pointers often point to. ReadOnly
collections are tested to behave well as long as they follow basic conventions. Many other features, such as cloning Dictionary
ies properly while keeping hashcodes, INotifyPropertyChanged
, delegate
s, event
s, HttpRequest
s / responses, and others are supported. If something doesn't work out of the box let me know in the issues, the repository is actively maintained.
FastCloner aims to work correctly and meet reasonable expectations by default while being fast. Benchmarking results are available here, check them out! By default, fast cloner relies on heavily cached reflection to work. An incremental source generator is currently in development as an opt-in alternative for performance-critical scenarios.
MIT, simple 💜