FlameCsv

High-performance RFC 4180-compliant CSV library for .NET 9 with trimming/AOT support

---

• Download on nuget
• See the documentation

Features

• TL;DR: Blazingly fast, trimmable and easy-to-use feature-rich CSV library
• 💡 Ease of Use
  • Simple API for reading and writing CSV
  • Built-in support for common CLR types
  • Supports both synchronous and asynchronous operations
  • Flexible; read or write almost any data source
  • Automatic newline detection
  • UTF-8/ASCII support to read/write bytes directly without additional transcoding
  • Supports hot reload
• 🚀 High Performance
  • Optimized for speed and low memory usage
  • SIMD-accelerated parsing routines with hardware intrinsics
  • Batteries-included internal caching and memory pooling for near-zero allocations
  • Reflection code paths that rival manually written code in performance
• 🛠️ Deep Customization
  • Read or write either .NET objects, or raw CSV records and fields
  • Attribute configuration for header names, constructors, field order, and more
  • Support for custom converters and converter factories
  • Read or write multiple CSV documents from/to a single data stream
• ✍️ Source Generator
  • Fully annotated and compatible with NativeAOT
  • Supports trimming to reduce application size
  • View and debug the code instead of opaque reflection
  • Compile-time diagnostics instead of runtime errors

Examples

Reading records

record User(int Id, string Name, DateTime LastLogin, int? Age = null);

string data = "id,name,lastlogin\n1,Bob,2010-01-01\n2,Alice,2024-05-22";

foreach (var user in CsvReader.Read<User>(data))
{
    Console.WriteLine(user);
}

Reflection-free reading using source generator

record User(int Id, string Name, DateTime LastLogin, int? Age = null);

[CsvTypeMap<char, User>]
partial class UserTypeMap;

foreach (var user in CsvReader.Read<User>(data, UserTypeMap.Default))
{
    Console.WriteLine(user);
}

Reading fields manually

string data = "id,name,lastlogin,age\n1,Bob,2010-01-01,42\n2,Alice,2024-05-22,\n";

CsvOptions<char> options = new()
{
    HasHeader = true,
    Comparer = StringComparer.Ordinal,
    Converters = { new CustomDateTimeConverter() },
};

foreach (CsvValueRecord<char> record in CsvReader.Enumerate(data, options))
{
    // get fields by column index or header name
    var u1 = new User(
        Id:        record.GetField<int>("Id"),
        Name:      record.GetField<string>("Name"),
        LastLogin: record.GetField<DateTime>("LastLogin"),
        Age:       record.FieldCount >= 3 ? record.GetField<int?>("Age") : null);

    var u2 = new User(
        Id:        record.GetField<int>(0),
        Name:      record.GetField<string>(1),
        LastLogin: record.GetField<DateTime>(2),
        Age:       record.FieldCount >= 3 ? record.GetField<int?>(3) : null);
}

Reading headerless CSV

class User
{
    [CsvIndex(0)] public int Id { get; set; }
    [CsvIndex(1)] public string? Name { get; set; }
    [CsvIndex(2)] public DateTime LastLogin { get; set; }
}

string data = "1,Bob,2010-01-01\n2,Alice,2024-05-22";

var options = new CsvOptions<char> { HasHeader = false };

foreach (var user in CsvReader.Read<User>(data, options))
{
    Console.WriteLine(user);
}

Writing records

User[] data = [new(1, "Bob", DateTime.Now, 42), new(2, "Alice", DateTime.UnixEpoch, null)];
StringBuilder result = CsvWriter.WriteToString(data);
Console.WriteLine(result);

Writing manually

using (CsvWriter<char> writer = CsvWriter.Create(TextWriter.Null))
{
    writer.WriteRecord(new User(1, "Bob", DateTime.Now, 42));
    writer.NextRecord();
    writer.Flush();

    writer.WriteField(1);
    writer.WriteField("Alice");
    writer.WriteField(DateTime.UnixEpoch);
    writer.WriteField((int?)null);
    writer.NextRecord();
}

Benchmarks

Reading 5000 records into objects

Method	Mean	Ratio	Allocated	Alloc Ratio
FlameCsv (Reflection)	2.308 ms	1.00	1.66 MB	1.00
FlameCsv (SourceGen)	2.506 ms	1.09	1.66 MB	1.00
Sylvan	2.570 ms	1.11	2.64 MB	1.59
RecordParser	4.673 ms	2.02	1.93 MB	1.16
CsvHelper	6.424 ms	2.78	3.49 MB	2.10

Iterating 65535 records without processing all fields

Method	Mean	Ratio	Allocated	Alloc Ratio
FlameCsv	3.292 ms	1.00	322 B	1.00
Sep	4.431 ms	1.35	5942 B	18.45
Sylvan	5.014 ms	1.52	42029 B	130.52
RecordParser	6.358 ms	1.93	2584418 B	8,026.14
CsvHelper	34.877 ms	10.60	2789195 B	8,662.10

Writing 5000 records

Method	Mean	Ratio	Allocated	Alloc Ratio
FlameCsv (SourceGen)	3.196 ms	1.00	170 B	1.00
FlameCsv (Reflection)	3.302 ms	1.03	174 B	1.02
Sylvan	3.467 ms	1.08	33605 B	197.68
Sep	3.561 ms	1.11	121181 B	712.83
CsvHelper	7.806 ms	2.44	2077347 B	12,219.69
RecordParser	9.245 ms	2.89	8691788 B	51,128.16