[XLSX] A cell should not contain more than 32,767 characters (#365)

src/Spout/Writer/XLSX/Internal/Worksheet.php

@@ -4,6 +4,7 @@ namespace Box\Spout\Writer\XLSX\Internal;
 
 use Box\Spout\Common\Exception\InvalidArgumentException;
 use Box\Spout\Common\Exception\IOException;
+use Box\Spout\Common\Helper\StringHelper;
 use Box\Spout\Writer\Common\Helper\CellHelper;
 use Box\Spout\Writer\Common\Internal\WorksheetInterface;
 
@@ -16,6 +17,14 @@ use Box\Spout\Writer\Common\Internal\WorksheetInterface;
  */
 class Worksheet implements WorksheetInterface
 {
+    /**
+     * Maximum number of characters a cell can contain
+     * @see https://support.office.com/en-us/article/Excel-specifications-and-limits-16c69c74-3d6a-4aaf-ba35-e6eb276e8eaa [Excel 2007]
+     * @see https://support.office.com/en-us/article/Excel-specifications-and-limits-1672b34d-7043-467e-8e27-269d656771c3 [Excel 2010]
+     * @see https://support.office.com/en-us/article/Excel-specifications-and-limits-ca36e2dc-1f09-4620-b726-67c00b05040f [Excel 2013/2016]
+     */
+    const MAX_CHARACTERS_PER_CELL = 32767;
+
     const SHEET_XML_FILE_HEADER = <<<EOD
 <?xml version="1.0" encoding="UTF-8" standalone="yes"?>
 <worksheet xmlns="http://schemas.openxmlformats.org/spreadsheetml/2006/main" xmlns:r="http://schemas.openxmlformats.org/officeDocument/2006/relationships">
@@ -39,6 +48,9 @@ EOD;
     /** @var \Box\Spout\Common\Escaper\XLSX Strings escaper */
     protected $stringsEscaper;
 
+    /** @var \Box\Spout\Common\Helper\StringHelper String helper */
+    protected $stringHelper;
+
     /** @var Resource Pointer to the sheet data file (e.g. xl/worksheets/sheet1.xml) */
     protected $sheetFilePointer;
 
@@ -62,6 +74,7 @@ EOD;
 
         /** @noinspection PhpUnnecessaryFullyQualifiedNameInspection */
         $this->stringsEscaper = \Box\Spout\Common\Escaper\XLSX::getInstance();
+        $this->stringHelper = new StringHelper();
 
         $this->worksheetFilePath = $worksheetFilesFolder . '/' . strtolower($this->externalSheet->getName()) . '.xml';
         $this->startSheet();
@@ -192,7 +205,7 @@ EOD;
      * @param mixed $cellValue
      * @param int $styleId
      * @return string
-     * @throws InvalidArgumentException
+     * @throws InvalidArgumentException If the given value cannot be processed
      */
     private function getCellXML($rowIndex, $cellNumber, $cellValue, $styleId)
     {
@@ -201,12 +214,7 @@ EOD;
         $cellXML .= ' s="' . $styleId . '"';
 
         if (CellHelper::isNonEmptyString($cellValue)) {
-            if ($this->shouldUseInlineStrings) {
+            $cellXML .= $this->getCellXMLFragmentForNonEmptyString($cellValue);
-                $cellXML .= ' t="inlineStr"><is><t>' . $this->stringsEscaper->escape($cellValue) . '</t></is></c>';
-            } else {
-                $sharedStringId = $this->sharedStringsHelper->writeString($cellValue);
-                $cellXML .= ' t="s"><v>' . $sharedStringId . '</v></c>';
-            }
         } else if (CellHelper::isBoolean($cellValue)) {
             $cellXML .= ' t="b"><v>' . intval($cellValue) . '</v></c>';
         } else if (CellHelper::isNumeric($cellValue)) {
@@ -226,6 +234,29 @@ EOD;
         return $cellXML;
     }
 
+    /**
+     * Returns the XML fragment for a cell containing a non empty string
+     *
+     * @param string $cellValue The cell value
+     * @return string The XML fragment representing the cell
+     * @throws InvalidArgumentException If the string exceeds the maximum number of characters allowed per cell
+     */
+    private function getCellXMLFragmentForNonEmptyString($cellValue)
+    {
+        if ($this->stringHelper->getStringLength($cellValue) > self::MAX_CHARACTERS_PER_CELL) {
+            throw new InvalidArgumentException('Trying to add a value that exceeds the maximum number of characters allowed in a cell (32,767)');
+        }
+
+        if ($this->shouldUseInlineStrings) {
+            $cellXMLFragment = ' t="inlineStr"><is><t>' . $this->stringsEscaper->escape($cellValue) . '</t></is></c>';
+        } else {
+            $sharedStringId = $this->sharedStringsHelper->writeString($cellValue);
+            $cellXMLFragment = ' t="s"><v>' . $sharedStringId . '</v></c>';
+        }
+
+        return $cellXMLFragment;
+    }
+
     /**
      * Closes the worksheet
      *

tests/Spout/Writer/XLSX/WriterTest.php

@@ -6,6 +6,7 @@ use Box\Spout\Common\Exception\SpoutException;
 use Box\Spout\Common\Type;
 use Box\Spout\TestUsingResource;
 use Box\Spout\Writer\WriterFactory;
+use Box\Spout\Writer\XLSX\Internal\Worksheet;
 
 /**
  * Class WriterTest
@@ -98,6 +99,19 @@ class WriterTest extends \PHPUnit_Framework_TestCase
     public function testAddRowShouldThrowExceptionIfUnsupportedDataTypePassedIn()
     {
         $fileName = 'test_add_row_should_throw_exception_if_unsupported_data_type_passed_in.xlsx';
+        $dataRows = [
+            [str_repeat('a', Worksheet::MAX_CHARACTERS_PER_CELL + 1)],
+        ];
+
+        $this->writeToXLSXFile($dataRows, $fileName);
+    }
+
+    /**
+     * @expectedException \Box\Spout\Common\Exception\InvalidArgumentException
+     */
+    public function testAddRowShouldThrowExceptionIfWritingStringExceedingMaxNumberOfCharactersAllowedPerCell()
+    {
+        $fileName = 'test_add_row_should_throw_exception_if_string_exceeds_max_num_chars_allowed_per_cell.xlsx';
         $dataRows = [
             [new \stdClass()],
         ];