Inline strings can contain multiple value nodes

We were working under the assumption that XLSX's inline strings only had a single value node (`<t>`). This is incorrect.
To get the actual value of an inline string node, we need to concatenate the value of all its child nodes.
This commit is contained in:
Adrien Loison 2021-05-14 16:10:12 +02:00
parent 69eeeff478
commit fde8a495ca
5 changed files with 30 additions and 7 deletions

View File

@ -14,7 +14,8 @@
"require": { "require": {
"php": ">=7.2.0", "php": ">=7.2.0",
"ext-zip": "*", "ext-zip": "*",
"ext-xmlreader" : "*" "ext-xmlreader": "*",
"ext-dom": "*"
}, },
"require-dev": { "require-dev": {
"phpunit/phpunit": "^8", "phpunit/phpunit": "^8",

View File

@ -122,10 +122,15 @@ class CellValueFormatter
*/ */
protected function formatInlineStringCellValue($node) protected function formatInlineStringCellValue($node)
{ {
// inline strings are formatted this way: // inline strings are formatted this way (they can contain any number of <t> nodes):
// <c r="A1" t="inlineStr"><is><t>[INLINE_STRING]</t></is></c> // <c r="A1" t="inlineStr"><is><t>[INLINE_STRING]</t><t>[INLINE_STRING_2]</t></is></c>
$tNode = $node->getElementsByTagName(self::XML_NODE_INLINE_STRING_VALUE)->item(0); $tNodes = $node->getElementsByTagName(self::XML_NODE_INLINE_STRING_VALUE);
$cellValue = $this->escaper->unescape($tNode->nodeValue);
$cellValue = '';
for ($i = 0; $i < $tNodes->count(); $i++) {
$tNode = $tNodes->item($i);
$cellValue .= $this->escaper->unescape($tNode->nodeValue);
}
return $cellValue; return $cellValue;
} }

View File

@ -186,18 +186,22 @@ class CellValueFormatterTest extends TestCase
public function testFormatInlineStringCellValue($value, $expectedFormattedValue) public function testFormatInlineStringCellValue($value, $expectedFormattedValue)
{ {
$nodeListMock = $this->createMock(\DOMNodeList::class); $nodeListMock = $this->createMock(\DOMNodeList::class);
$nodeListMock
->expects($this->atLeastOnce())
->method('count')
->willReturn(1);
$nodeListMock $nodeListMock
->expects($this->atLeastOnce()) ->expects($this->atLeastOnce())
->method('item') ->method('item')
->with(0) ->with(0)
->will($this->returnValue((object) ['nodeValue' => $value])); ->willReturn((object) ['nodeValue' => $value]);
$nodeMock = $this->createMock(\DOMElement::class); $nodeMock = $this->createMock(\DOMElement::class);
$nodeMock $nodeMock
->expects($this->atLeastOnce()) ->expects($this->atLeastOnce())
->method('getElementsByTagName') ->method('getElementsByTagName')
->with(CellValueFormatter::XML_NODE_INLINE_STRING_VALUE) ->with(CellValueFormatter::XML_NODE_INLINE_STRING_VALUE)
->will($this->returnValue($nodeListMock)); ->willReturn($nodeListMock);
$formatter = new CellValueFormatter(null, null, false, false, new Escaper\XLSX()); $formatter = new CellValueFormatter(null, null, false, false, new Escaper\XLSX());
$formattedValue = \ReflectionHelper::callMethodOnObject($formatter, 'formatInlineStringCellValue', $nodeMock); $formattedValue = \ReflectionHelper::callMethodOnObject($formatter, 'formatInlineStringCellValue', $nodeMock);

View File

@ -72,6 +72,19 @@ class ReaderTest extends TestCase
} }
} }
/**
* @return void
*/
public function testReadShouldSupportInlineStringsWithMultipleValueNodes()
{
$allRows = $this->getAllRowsForFile('sheet_with_multiple_value_nodes_in_inline_strings.xlsx');
$expectedRows = [
['VALUE 1 VALUE 2 VALUE 3 VALUE 4', 's1 - B1'],
];
$this->assertEquals($expectedRows, $allRows);
}
/** /**
* @return void * @return void
*/ */