Performance · benchmark reproductible

4.5× faster than Spring Batch Java

10 million financial transactions · CSV → PostgreSQL → XML · same chunk size, same connection pool

Total pipeline

42s

4.5× faster · 187s

Peak memory

30× less · 62 MB vs 1 840 MB

Cold start

320× faster · <10ms vs 3.2s

Spring Batch RS (Rust)Spring Batch (Java)

→ Methodology, full numbers & how to reproduce

LIVE EXAMPLE

CSV → JSON Pipeline

Type-safe, fault-tolerant, zero boilerplate

use spring_batch_rs::prelude::*;

// 1. Define your data shape — Rust enforces type safety at compile time
#[derive(Deserialize, Serialize)]
struct Product {
    id: u32,
    name: String,
    price: f64,
}

// 2. Configure source and destination using builder patterns
let reader = CsvItemReaderBuilder::<Product>::new()
    .from_path("products.csv")
    .has_headers(true)
    .build();

// Passthrough processor — forwards items unchanged (swap in your own logic here)
let processor = PassThroughProcessor::<Product>::new();

let writer = JsonItemWriterBuilder::<Product>::new()
    .from_path("products.json");

// 3. Build the pipeline
//    chunk(100)      → accumulate 100 items, then write once (balances I/O vs memory)
//    skip_limit(10)  → tolerate up to 10 bad records before the job fails
let step = StepBuilder::new("convert-products")
    .reader(&reader)
    .processor(&processor)
    .writer(&writer)
    .chunk(100)
    .skip_limit(10)
    .build();

// 4. Run — the job tracks status, counts, and errors automatically
JobBuilder::new().start(&step).build().run();

→ View full source on GitHub

Getting Started First batch job in 5 minutes. Installation, setup, first example.

Java vs Rust Benchmark Methodology, full numbers & how to reproduce on your own infrastructure.

Examples Gallery 24+ pipelines with links to source code on GitHub.

Architecture Job → Step → Reader/Processor/Writer. Chunk-oriented processing.