Fixes parsing CSV files with BOM

https://en.wikipedia.org/wiki/Byte_order_mark
webgriffe · Feb 15, 2019 · 825a276 · 825a276
1 parent 00d725f
commit 825a276
Show file tree

Hide file tree

Showing 4 changed files with 43 additions and 2 deletions.
diff --git a/src/Iterator.php b/src/Iterator.php
@@ -83,8 +83,11 @@ private function attachEmitter()
                         $this->emitter->fail(
                             new \LogicException(
                                 sprintf(
-                                    'Invalid number of columns at line %d of given CSV file.',
-                                    $this->csvParser->getRowsParsed()
+                                    'Invalid number of columns at line %d of given CSV file. Header has %d columns, ' .
+                                    'this line %d columns.',
+                                    $this->csvParser->getRowsParsed(),
+                                    \count($header),
+                                    \count($row)
                                 )
                             )
                         );

diff --git a/src/Parser.php b/src/Parser.php
@@ -44,12 +44,17 @@ public function __construct(
     public function parseRow(): Promise
     {
         return call(function () {
+            $isFirstRead = $this->fileHandle->tell() === 0;
             if ($this->fileHandle->eof()) {
                 return null;
             }
             $buffer = '';
             $newLinePos = null;
             while ($chunk = yield $this->fileHandle->read()) {
+                if ($isFirstRead) {
+                    $chunk = $this->removeBom($chunk);
+                }
+                $isFirstRead = false;
                 $buffer .= $chunk;
                 $newLinePos = strpos($buffer, PHP_EOL);
                 if ($newLinePos !== false) {
@@ -85,4 +90,17 @@ public function getRowsParsed(): int
     {
         return $this->rowsParsed;
     }
+
+    /**
+     * @param $chunk
+     * @return string
+     */
+    private function removeBom(string $chunk): string
+    {
+        $bom = pack('CCC', 0xEF, 0xBB, 0xBF);
+        if (strpos($chunk, $bom) === 0) {
+            $chunk = (string)substr($chunk, 3);
+        }
+        return $chunk;
+    }
 }
diff --git a/tests/ParserTest.php b/tests/ParserTest.php
@@ -164,4 +164,21 @@ public function testParseFileWithEscapedCharactersInValues()
         $this->assertEquals('Algeria', $rows[2][0]);
         $this->assertEquals($expectedAlgeriaDescription, $rows[2][1]);
     }
+
+    public function testParseFileWithBom()
+    {
+        $rows = [];
+        Loop::run(function () use (&$rows) {
+            $parser = new Parser(yield File\open(__DIR__ . '/file-with-bom.csv', 'rb'), ';');
+            while ($row = yield $parser->parseRow()) {
+                $rows[] = $row;
+            }
+        });
+        $this->assertCount(3, $rows);
+        $this->assertEquals(
+            'sku',
+            $rows[0][0],
+            'It appears that there is a BOM because the first field of the first row doesn\'t match the expected value.'
+        );
+    }
 }
diff --git a/tests/file-with-bom.csv b/tests/file-with-bom.csv
@@ -0,0 +1,3 @@
+sku;name
+1;first
+2;second