README

A modern, typed, and lazy PHP 8.4+ data pipeline library.

php-etl is a high-performance ETL (Extract, Transform, Load) library designed to handle massive datasets with a constant memory footprint. Using PHP Generators under the hood, it processes data row-by-row, ensuring even the largest pipelines remain memory-safe.

Features

• Lazy Evaluation: Entirely stream-based using Generators. Never materialize large arrays again.
• Typed Pipelines: Full support for PHPStan generics (Pipeline<TIn, TOut>).
• High-Performance Extractors: Built-in support for CSV and JSON Lines (ndjson) with zero memory overhead.
• Batching & Chunking: Efficiently process rows in batches using chunk($size) or batch($size).
• Fluent API: Modern entry point via Pipeline::create().
• Error Resilience: Configurable error policies (COLLECT, SKIP, THROW) to manage row-level failures.
• Bulk Loading: Built-in BulkPdoLoader for high-performance database inserts.
• Zero Dependencies: Lightweight core with no runtime dependencies.

Installation

composer require fr3on/php-etl

Usage

Basic Example

use Fr3on\Etl\Pipeline;
use Fr3on\Etl\Extractor\CsvExtractor;
use Fr3on\Etl\Sink\CallbackSink;

Pipeline::create()
    ->extract(new CsvExtractor('transactions.csv'))
    ->filter(fn($row) => $row['status'] === 'SUCCESS')
    ->map(fn($row) => ['id' => $row['tx_id'], 'amt' => (float)$row['amount']])
    ->into(new CallbackSink(fn($row) => print_r($row)))
    ->run();

Bulk Loading Example

use Fr3on\Etl\Loader\BulkPdoLoader;

Pipeline::create()
    ->extract(new CsvExtractor('large_data.csv'))
    ->chunk(1000)
    ->load(new BulkPdoLoader($pdo, 'target_table'))
    ->run();

Performance

php-etl is built for extreme efficiency. Benchmarks on 100,000 rows demonstrate sub-second processing with a constant memory footprint.

Benchmarks run on standard hardware (PHP 8.4, Opcache off). Results may vary depending on row complexity.

Testing & Benchmarking

The core library is tested for memory safety and high throughput.

composer test         # Run PHPUnit
composer test:memory  # Run with strict 32M memory limit
composer analyse      # Run PHPStan Level 9
composer bench        # Run PHPBench

License

MIT License. See LICENSE for details.