Introduction

🔥 FlameCSV 🔥 is a fully-featured high performance CSV library for .NET with a simple API, deep customization options, and built-in support for UTF8, nativeAOT, and more.

FlameCsv is consistently the fastest .NET CSV library in the available benchmarks.

Read CSV at multiple gigabytes per second on consumer hardware, and write arbitrarily large amounts of CSV with near-zero allocations. FlameCSV leverages modern .NET patterns and libraries such as spans, SIMD hardware intrinsics, memory/string pooling, pipes and buffer writers, and is built from the ground up to provide an easy-to-use high performance experience.

The library has thousands of tests, and critical paths have been fuzz-tested with SharpFuzz.

See Getting Started, view the Examples, or deep dive into the API Reference.

Features

💡 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
- Specialized SIMD-accelerated parsing and unescaping
- Batteries-included internal caching and memory pooling for near-zero allocations
- Reflection code paths that rival and exceed 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
- Debuggable source code instead of compiled expressions
- Compile-time diagnostics instead of runtime errors
- Feature parity with reflection-based code paths (soon)

Example

UTF-16
UTF-8

using FlameCsv;

const string data =
    """
    id,name,age
    1,Bob,42
    2,Alice,37
    3,"Bond, James",39
    """;

List<User> users = [];

// read users from utf16 string
foreach (var user in CsvReader.Read<User>(data))
{
    users.Add(user);
}

// write users to a stream as tab-separated fields
await CsvWriter.WriteAsync(
    new StreamWriter(stream, Encoding.UTF8),
    users,
    new CsvOptions<char> { Delimiter = '\t' },
    cancellationToken);

using FlameCsv;

byte[] data =
    """
    id,name,age
    1,Bob,42
    2,Alice,37
    3,"Bond, James",39
    """u8.ToArray();

List<User> users = [];

// read users from utf8 bytes
foreach (var user in CsvReader.Read<User>(data))
{
    users.Add(user);
}

// write users to a stream as tab-separated fields
await CsvWriter.WriteAsync(
    stream,
    users,
    new CsvOptions<byte> { Delimiter = '\t' },
    cancellationToken);

Dependencies

FlameCsv has two dependencies:

System.IO.Pipelines for the powerful abstractions and additional @"FlameCsv.CsvReader.ReadAsync``1(System.IO.Pipelines.PipeReader,FlameCsv.Binding.CsvTypeMap{System.Byte,``0},FlameCsv.CsvOptions{System.Byte})?text=reading" and writing APIs
CommunityToolkit.HighPerformance which provides utilities for writing high-performance code, and the string pool used for header values.

A development-time dependency to FastExpressionCompiler is used for extremely performant runtime code-generation in reflection-based code paths.

It's possible in the future to split FlameCsv into separate packages to keep the core library dependency-free, but there are no hard plans for it yet.