Specify data type in html tags using attributes PHPOffice#3444

PouriaSeyfi · Mar 10, 2023 · eac0aef · eac0aef
1 parent 66221bf
commit eac0aef
Show file tree

Hide file tree

Showing 2 changed files with 41 additions and 0 deletions.
diff --git a/src/PhpSpreadsheet/Reader/Html.php b/src/PhpSpreadsheet/Reader/Html.php
@@ -564,6 +564,18 @@ private function processDomElementDataFormat(Worksheet $sheet, int $row, string
         }
     }
 
+    private function processDomElementDataType(Worksheet $sheet, int $row, string $column, array $attributeArray): void
+    {
+        if (isset($attributeArray['data-type'])) {
+            $cell = $sheet->getCell($column . $row);
+            $cellValue = $cell->getValue();
+            //cast value to the datatype
+            $sheet->setCellValueExplicit($column . $row, $cellValue, $attributeArray['data-type']);
+            //set datatype
+            $cell->setDataType($attributeArray['data-type']);
+        }
+    }
+
     private function processDomElementThTd(Worksheet $sheet, int &$row, string &$column, string &$cellContent, DOMElement $child, array &$attributeArray): void
     {
         while (isset($this->rowspan[$column . $row])) {
@@ -582,6 +594,7 @@ private function processDomElementThTd(Worksheet $sheet, int &$row, string &$col
         $this->processDomElementAlign($sheet, $row, $column, $attributeArray);
         $this->processDomElementVAlign($sheet, $row, $column, $attributeArray);
         $this->processDomElementDataFormat($sheet, $row, $column, $attributeArray);
+        $this->processDomElementDataType($sheet, $row, $column, $attributeArray);
 
         if (isset($attributeArray['rowspan'], $attributeArray['colspan'])) {
             //create merging rowspan and colspan

diff --git a/tests/PhpSpreadsheetTests/Reader/Html/HtmlTest.php b/tests/PhpSpreadsheetTests/Reader/Html/HtmlTest.php
@@ -2,6 +2,7 @@
 
 namespace PhpOffice\PhpSpreadsheetTests\Reader\Html;
 
+use PhpOffice\PhpSpreadsheet\Cell\DataType;
 use PhpOffice\PhpSpreadsheet\Reader\Exception as ReaderException;
 use PhpOffice\PhpSpreadsheet\Reader\Html;
 use PhpOffice\PhpSpreadsheet\Style\Alignment;
@@ -363,4 +364,31 @@ public function testBorderWithColspan(): void
             self::assertEquals(Border::BORDER_THIN, $border->getBorderStyle());
         }
     }
+
+    public function testDataType(): void
+    {
+        $html = '<table>
+                    <tr>
+                        <td data-type="b">1</td>
+                        <td data-type="s">12345678987654321</td>
+                        <td data-type="f">=CONCAT("TEXT A ","TEXT B")</td>
+                    </tr>
+                </table>';
+        $reader = new Html();
+        $spreadsheet = $reader->loadFromString($html);
+        $firstSheet = $spreadsheet->getSheet(0);
+
+        // check boolean data type
+        self::assertEquals(DataType::TYPE_BOOL, $firstSheet->getCell('A1')->getDataType());
+        self::assertIsBool($firstSheet->getCell('A1')->getValue());
+
+        // check string data type
+        self::assertEquals(DataType::TYPE_STRING, $firstSheet->getCell('B1')->getDataType());
+        self::assertIsString($firstSheet->getCell('B1')->getValue());
+
+        // check formula data type
+        self::assertEquals(DataType::TYPE_FORMULA, $firstSheet->getCell('C1')->getDataType());
+        // check formula output
+        self::assertEquals('TEXT A TEXT B', $firstSheet->getCell('C1')->getFormattedValue());
+    }
 }