Merge pull request #5 from clue-labs/ndjson

Example and documentation for converting CSV to NDJSON

Author: clue

README.md

@@ -105,6 +105,15 @@ that CSV is often used for historial reasons and using CSV to store structured
 application data is usually not a good idea nowadays – but exporting to CSV for
 known applications is a very reasonable approach.
 
+As an alternative, if you want to process structured data in a more modern
+JSON-based format, you may want to use [clue/reactphp-ndjson](https://github.com/clue/reactphp-ndjson)
+to process newline-delimited JSON (NDJSON) files (`.ndjson` file extension).
+
+```json
+{"name":"Alice","age":30,"comment":"Yes, I like cheese"}
+{"name":"Bob","age":50,"comment":"Hello\nWorld!"}
+```
+
 ## Usage
 
 ### Decoder
@@ -329,6 +338,10 @@ This project is released under the permissive [MIT license](LICENSE).
 * If you want to learn more about processing streams of data, refer to the documentation of
   the underlying [react/stream](https://github.com/reactphp/stream) component.
 
+* If you want to process structured data in a more modern JSON-based format,
+  you may want to use [clue/reactphp-ndjson](https://github.com/clue/reactphp-ndjson)
+  to process newline-delimited JSON (NDJSON) files (`.ndjson` file extension).
+
 * If you want to process compressed CSV files (`.csv.gz` file extension)
   you may want to use [clue/reactphp-zlib](https://github.com/clue/reactphp-zlib)
   on the compressed input stream before passing the decompressed stream to the CSV decoder.

examples/11-csv2ndjson.php

@@ -0,0 +1,45 @@
+<?php
+
+use React\EventLoop\Factory;
+use React\Stream\ReadableResourceStream;
+use React\Stream\WritableResourceStream;
+use Clue\React\Csv\Decoder;
+use React\Stream\ThroughStream;
+
+require __DIR__ . '/../vendor/autoload.php';
+
+$loop = Factory::create();
+
+$exit = 0;
+$in = new ReadableResourceStream(STDIN, $loop);
+$out = new WritableResourceStream(STDOUT, $loop);
+$info = new WritableResourceStream(STDERR, $loop);
+
+$delimiter = isset($argv[1]) ? $argv[1] : ',';
+
+$decoder = new Decoder($in, $delimiter);
+
+$headers = array();
+$encoder = new ThroughStream(function ($data) use (&$headers) {
+    return json_encode(array_combine($headers, $data), JSON_UNESCAPED_SLASHES | JSON_UNESCAPED_UNICODE) . "\n";
+});
+$encoder->pipe($out);
+
+// first row from decoder will be used as header values, then start piping to encoder
+$decoder->once('data', function ($data) use (&$headers, $decoder, $encoder) {
+    $headers = $data;
+    $decoder->pipe($encoder);
+});
+
+$decoder->on('error', function (Exception $e) use ($info, &$exit) {
+    $info->write('ERROR: ' . $e->getMessage() . PHP_EOL);
+    $exit = 1;
+});
+
+$info->write('You can pipe/write a valid CSV stream to STDIN' . PHP_EOL);
+$info->write('Valid NDJSON (Newline-Delimited JSON) will be forwarded to STDOUT' . PHP_EOL);
+$info->write('Invalid CSV will raise an error on STDERR and exit with code 1' . PHP_EOL);
+
+$loop->run();
+
+exit($exit);