PHPackages sanix-darker/cisv - PHPackages - PHPackages [Skip to content](#main-content)[PHPackages](/)[Directory](/)[Categories](/categories)[Trending](/trending)[Leaderboard](/leaderboard)[Changelog](/changelog)[Analyze](/analyze)[Collections](/collections)[Log in](/login)[Sign up](/register) 1. [Directory](/) 2. / 3. [PDF & Document Generation](/categories/documents) 4. / 5. sanix-darker/cisv ArchivedPhp-ext[PDF & Document Generation](/categories/documents) sanix-darker/cisv ================= High-performance CSV parser with SIMD optimizations (AVX-512/AVX2) for PHP - 10-50x faster than fgetcsv v0.4.8(2mo ago)216[9 issues](https://github.com/Sanix-Darker/cisv-php/issues)MITCPHP >=8.1CI passing Since Mar 6Pushed 2mo agoCompare [ Source](https://github.com/Sanix-Darker/cisv-php)[ Packagist](https://packagist.org/packages/sanix-darker/cisv)[ Docs](https://github.com/Sanix-Darker/cisv-php)[ RSS](/packages/sanix-darker-cisv/feed)WikiDiscussions master Synced 1mo ago READMEChangelog (10)Dependencies (4)Versions (70)Used By (0) CISV ==== [](#cisv) [![License](https://camo.githubusercontent.com/b8cadaa967891081f8f165695470689986c028821dd8a040132f6e661795dc0d/68747470733a2f2f696d672e736869656c64732e696f2f62616467652f6c6963656e73652d4d49542d626c7565)](https://camo.githubusercontent.com/b8cadaa967891081f8f165695470689986c028821dd8a040132f6e661795dc0d/68747470733a2f2f696d672e736869656c64732e696f2f62616467652f6c6963656e73652d4d49542d626c7565)[![Build](https://camo.githubusercontent.com/b0c6c6845a74cb65a7f0a32bdcfd8fbf80eeb40026c4029af424ab371c94b8bd/68747470733a2f2f696d672e736869656c64732e696f2f62616467652f6275696c642d70617373696e672d627269676874677265656e)](https://camo.githubusercontent.com/b0c6c6845a74cb65a7f0a32bdcfd8fbf80eeb40026c4029af424ab371c94b8bd/68747470733a2f2f696d672e736869656c64732e696f2f62616467652f6275696c642d70617373696e672d627269676874677265656e)[![Ask DeepWiki](https://camo.githubusercontent.com/0f5ae213ac378635adeb5d7f13cef055ad2f7d9a47b36de7b1c67dbe09f609ca/68747470733a2f2f6465657077696b692e636f6d2f62616467652e737667)](https://deepwiki.com/Sanix-Darker/cisv) DEPRECATED REPOSITORY --------------------- [](#deprecated-repository) This monorepo is deprecated and no longer the primary source for installs/releases. Use the split repositories instead: - Core C library: - CLI: - PHP binding: - Node.js binding: - Python binding: High-performance CSV parser and writer with SIMD optimizations (AVX-512/AVX2). Available as a C library, CLI tool, and bindings for Node.js, Python, and PHP. Blog post: [How I accidentally created the fastest CSV parser ever made](https://sanixdk.xyz/blogs/how-i-accidentally-created-the-fastest-csv-parser-ever-made) FEATURES -------- [](#features) - SIMD-accelerated parsing (AVX-512, AVX2, with scalar fallback) - Zero-copy memory-mapped file processing - Streaming API for large files - Row-by-row iterator API (fgetcsv-style) with early exit support - Configurable delimiters, quotes, escapes, and comments - Field transforms (uppercase, lowercase, trim, hash, base64, etc.) - Parallel chunk processing for multi-threaded parsing - RFC 4180 compliant with relaxed mode option - CSV writer with SIMD optimization PERFORMANCE ----------- [](#performance) Benchmarks on Intel Core i5-8365U: TestThroughputSimple CSV165-290 MB/sQuoted fields175-189 MB/sLarge fields153-169 MB/sBenchmark reports are generated by `scripts/benchmark_cli.sh` in CI and committed per binding: - `BENCHMARKS_CLI.md` - `BENCHMARKS_NODEJS.md` - `BENCHMARKS_PYTHON.md` - `BENCHMARKS_PHP.md` INSTALLATION ------------ [](#installation) ### BUILD FROM SOURCE [](#build-from-source) ``` git clone https://github.com/sanix-darker/cisv cd cisv make all ``` ### INSTALL CLI [](#install-cli) ``` sudo make install-cli ``` ### NODE.JS [](#nodejs) ``` cd bindings/nodejs npm install npm run build ``` ### PYTHON [](#python) ``` cd bindings/python pip install -e . ``` ### PHP [](#php) ``` cd bindings/php phpize ./configure make sudo make install ``` CLI USAGE --------- [](#cli-usage) ``` cisv [OPTIONS] FILE Options: -d, --delimiter CHAR Field delimiter (default: ,) -q, --quote CHAR Quote character (default: ") -e, --escape CHAR Escape character -m, --comment CHAR Comment line prefix -t, --trim Trim whitespace from fields -r, --relaxed Relaxed parsing mode -c, --count Count rows only -s, --select COLS Select columns by index -o, --output FILE Output file -b, --benchmark Benchmark mode --skip-empty Skip empty lines --skip-errors Skip lines with errors --from-line N Start at line N --to-line N Stop at line N --head N First N rows --tail N Last N rows ``` ### EXAMPLES [](#examples) ``` # Parse CSV cisv data.csv # Parse TSV cisv -d $'\t' data.tsv # Count rows cisv -c data.csv # Select columns 0, 2, 5 cisv -s 0,2,5 data.csv # Benchmark cisv -b data.csv # Parse lines 100-500 cisv --from-line 100 --to-line 500 data.csv ``` ### CSV WRITER [](#csv-writer) ``` cisv write [OPTIONS] Options: -g, --generate N Generate N test rows -o, --output FILE Output file -d, --delimiter CHAR Field delimiter -Q, --always-quote Quote all fields -r, --crlf Use CRLF line endings -n, --null TEXT NULL representation ``` DOCKER BENCHMARKS ----------------- [](#docker-benchmarks) ``` docker build -t cisv-benchmark . docker run --rm --cpus="2.0" --memory="4g" cisv-benchmark ``` Per-component benchmark environments (useful when local tools are missing): ``` # CLI tools benchmark docker build -t cisv-cli-bench -f cli/benchmarks/Dockerfile . docker run --rm --platform linux/amd64 --cpus=2 --memory=4g cisv-cli-bench # Node.js benchmark docker build -t cisv-node-bench -f bindings/nodejs/benchmarks/Dockerfile . docker run --rm --platform linux/amd64 --cpus=2 --memory=4g cisv-node-bench # PHP benchmark docker build -t cisv-php-bench -f bindings/php/benchmarks/Dockerfile . docker run --rm --platform linux/amd64 --cpus=2 --memory=4g cisv-php-bench # Python (nanobind) benchmark docker build -t cisv-pynb-bench -f bindings/python-nanobind/benchmarks/Dockerfile . docker run --rm --platform linux/amd64 --cpus=2 --memory=4g cisv-pynb-bench ``` C API ----- [](#c-api) ``` #include #include #include // Callbacks void on_field(void *user, const char *data, size_t len); void on_row(void *user); // Parse file cisv_config cfg; cisv_config_init(&cfg); cfg.field_cb = on_field; cfg.row_cb = on_row; cfg.delimiter = ','; cisv_parser *p = cisv_parser_create_with_config(&cfg); cisv_parser_parse_file(p, "data.csv"); cisv_parser_destroy(p); // Count rows (fast mode) size_t count = cisv_parser_count_rows("data.csv"); // Row-by-row iterator (fgetcsv-style, supports early exit) cisv_iterator_t *it = cisv_iterator_open("data.csv", &cfg); const char **fields; const size_t *lengths; size_t field_count; while (cisv_iterator_next(it, &fields, &lengths, &field_count) == CISV_ITER_OK) { // Process row - fields valid until next call if (should_stop) break; // Early exit supported } cisv_iterator_close(it); // Parallel processing cisv_mmap_file_t *file = cisv_mmap_open("data.csv"); int chunk_count; cisv_chunk_t *chunks = cisv_split_chunks(file, 4, &chunk_count); // Parse chunks in parallel threads cisv_mmap_close(file); ``` NODE.JS API ----------- [](#nodejs-api) ``` const { cisvParser } = require('cisv'); // Parse file const parser = new cisvParser({ delimiter: ',', quote: '"', trim: true }); const rows = parser.parseSync('data.csv'); // Parse with transforms parser.transform(0, 'uppercase') .transform(1, 'trim') .transform(2, (value) => value.toUpperCase()); const transformedRows = parser.parseSync('data.csv'); // Name-based transforms parser.setHeaderFields(['id', 'name', 'email']); parser.transformByName('name', 'uppercase'); // Stream parsing const streamParser = new cisvParser(); streamParser.write(chunk); streamParser.end(); const results = streamParser.getRows(); // Row-by-row iteration (memory efficient, supports early exit) const iterParser = new cisvParser({ delimiter: ',' }); iterParser.openIterator('large.csv'); let row; while ((row = iterParser.fetchRow()) !== null) { console.log(row); if (row[0] === 'stop') break; // Early exit } iterParser.closeIterator(); // Fast row counting const total = cisvParser.countRows('large.csv'); ``` PYTHON API ---------- [](#python-api) ``` import cisv # Parse file rows = cisv.parse_file('data.csv', delimiter=',', trim=True) # Parse string rows = cisv.parse_string('a,b,c\n1,2,3') # Count rows (fast) count = cisv.count_rows('data.csv') # Row-by-row iteration (memory efficient, supports early exit) with cisv.CisvIterator('large.csv') as reader: for row in reader: print(row) if row[0] == 'stop': break # Early exit - no wasted work # Or use the convenience function for row in cisv.open_iterator('data.csv'): process(row) ``` PHP API ------- [](#php-api) ```