Fix for malformed tables

Webnuvola · Feb 23, 2021 · de1af45 · de1af45
1 parent a1f0efb
commit de1af45
Show file tree

Hide file tree

Showing 6 changed files with 61 additions and 8 deletions.
diff --git a/CHANGELOG.md b/CHANGELOG.md
@@ -1,6 +1,9 @@
 # Changelog
 All notable changes to `excel-reader` will be documented in this file.
 
+## v1.1.1 - 2021-02-23
+- Fix for malformed tables
+
 ## v1.1.0 - 2021-02-23
 - `ExcelReader` now requires `FileInterface` instead of `string` in `__construct`
 - Deprecated `ExcelReader::createFromFile` in favor of `ExcelReader::createFromPath`

diff --git a/src/Libraries/BoxSpout2Library.php b/src/Libraries/BoxSpout2Library.php
@@ -30,6 +30,8 @@ public function read(string $path, bool $hasHeaders, $sheetId): array
         $function = is_int($sheetId) ? 'getIndex' : 'getName';
 
         $headers = [];
+        $headersCount = 0;
+        $filler = [];
         $data = [];
 
         $slugify = $hasHeaders ? new Slugify() : null;
@@ -47,11 +49,20 @@ public function read(string $path, bool $hasHeaders, $sheetId): array
                 if ($first) {
                     $headers = array_map(static fn ($cell) => $slugify->slugify($cell), $row);
 
+                    $headersCount = count($headers);
+                    $filler = array_fill(0, $headersCount, null);
+
                     $first = false;
                     continue;
                 }
 
-                $data[] = $row;
+                $currentData = $row + $filler;
+
+                if ($hasHeaders && count($currentData) > $headersCount) {
+                    $currentData = array_slice($currentData, 0, $headersCount);
+                }
+
+                $data[] = $currentData;
             }
 
             break;

diff --git a/src/Libraries/BoxSpout3Library.php b/src/Libraries/BoxSpout3Library.php
@@ -28,14 +28,17 @@ public function read(string $path, bool $hasHeaders, $sheetId): array
         $function = is_int($sheetId) ? 'getIndex' : 'getName';
 
         $headers = [];
+        $headersCount = 0;
+        $filler = [];
         $data = [];
 
+        $slugify = $hasHeaders ? new Slugify() : null;
+
         foreach ($reader->getSheetIterator() as $sheet) {
             if ($sheet->$function() !== $sheetId) {
                 continue;
             }
 
-            $slugify = new Slugify();
             $first = $hasHeaders;
 
             /** @var \Box\Spout\Common\Entity\Row $row */
@@ -47,14 +50,23 @@ public function read(string $path, bool $hasHeaders, $sheetId): array
                         $row->getCells(),
                     );
 
+                    $headersCount = count($headers);
+                    $filler = array_fill(0, $headersCount, null);
+
                     $first = false;
                     continue;
                 }
 
-                $data[] = array_map(
-                    static fn (Cell $cell) => $cell->getValue(),
-                    $row->getCells(),
-                );
+                $currentData = array_map(
+                        static fn (Cell $cell) => $cell->getValue(),
+                        $row->getCells(),
+                    ) + $filler;
+
+                if ($hasHeaders && count($currentData) > $headersCount) {
+                    $currentData = array_slice($currentData, 0, $headersCount);
+                }
+
+                $data[] = $currentData;
             }
 
             break;

diff --git a/tests/Unit/ExcelReaderTest.php b/tests/Unit/ExcelReaderTest.php
@@ -3,6 +3,7 @@
 namespace Webnuvola\ExcelReader\Tests\Unit;
 
 use DateTime;
+use RuntimeException;
 use Webnuvola\ExcelReader\ExcelReader;
 use Webnuvola\ExcelReader\ExcelReaderManager;
 use Webnuvola\ExcelReader\Libraries\BoxSpout2Library;
@@ -134,4 +135,30 @@ public function read_file_from_string()
 
         $this->assertCount(700, $excel);
     }
+
+    /** @test */
+    public function malformed_table_with_headers()
+    {
+        $excel = ExcelReader::createFromPath(__DIR__.'/../resources/malformed-table.xlsx')
+            ->read();
+
+        $this->assertEquals([
+            ['column-a' => 'Value A1', 'column-b' => 'Value B1', 'column-c' => 'Value C1', 'column-d' => 'Value D1'],
+            ['column-a' => 'Value A2', 'column-b' => 'Value B2', 'column-c' => 'Value C2', 'column-d' => null],
+        ], $excel);
+    }
+
+    /** @test */
+    public function malformed_table_without_headers()
+    {
+        $excel = ExcelReader::createFromPath(__DIR__.'/../resources/malformed-table.xlsx')
+            ->withoutHeaders()
+            ->read();
+
+        $this->assertEquals([
+            ['Column A', 'Column B', 'Column C', 'Column D'],
+            ['Value A1', 'Value B1', 'Value C1', 'Value D1', 'Value E1'],
+            ['Value A2', 'Value B2', 'Value C2'],
+        ], $excel);
+    }
 }
diff --git a/tests/Unit/FileTest.php b/tests/Unit/FileTest.php
@@ -41,9 +41,9 @@ public function temporary_file_is_deleted()
         $file = FileFactory::createFromString(self::$excelContent, 'xlsx');
 
         $path = $file->getPath();
-        $this->assertTrue(file_exists($path));
+        $this->assertFileExists($path);
 
         unset($file);
-        $this->assertFalse(file_exists($path));
+        $this->assertFileDoesNotExist($path);
     }
 }
diff --git a/tests/resources/malformed-table.xlsx b/tests/resources/malformed-table.xlsx