このスレッドで見つかったコードに基づいて単純なクラスを適応させ、現在は と を処理colspan
していrowspan
ます。十分にテストされておらず、最適化できると確信しています。
使用法:
require_once('table2csv.php');
$table = '<table border="1">
<tr>
<th colspan=2>Header 1</th>
</tr>
<tr>
<td>row 1, cell 1</td>
<td>row 1, cell 2</td>
</tr>
<tr>
<td>row 2, cell 1</td>
<td>row 2, cell 2</td>
</tr>
<tr>
<td rowspan=2>top left row</td>
<td>top right row</td>
</tr>
<tr>
<td>bottom right</td>
</tr>
</table>';
table2csv($table,"sample.csv",true);
table2csv.php
<?php
//download @ http://simplehtmldom.sourceforge.net/
require_once('simple_html_dom.php');
$repeatContentIntoSpannedCells = false;
//--------------------------------------------------------------------------------------------------------------------
function table2csv($rawHTML,$filename,$repeatContent) {
//get rid of sups - they mess up the wmus
for ($i=1; $i <= 20; $i++) {
$rawHTML = str_replace("<sup>".$i."</sup>", "", $rawHTML);
}
global $repeatContentIntoSpannedCells;
$html = str_get_html(trim($rawHTML));
$repeatContentIntoSpannedCells = $repeatContent;
//we need to pre-initialize the array based on the size of the table (how many rows vs how many columns)
//counting rows is easy
$rowCount = count($html->find('tr'));
//column counting is a bit trickier, we have to iterate through the rows and basically pull out the max found
$colCount = 0;
foreach ($html->find('tr') as $element) {
$tempColCount = 0;
foreach ($element->find('th') as $cell) {
$tempColCount++;
}
if ($tempColCount == 0) {
foreach ($element->find('td') as $cell) {
$tempColCount++;
}
}
if ($tempColCount > $colCount) $colCount = $tempColCount;
}
$mdTable = array();
for ($i=0; $i < $rowCount; $i++) {
array_push($mdTable, array_fill(0, $colCount, NULL));
}
//////////done predefining array
$rowPos = 0;
$fp = fopen($filename, "w");
foreach ($html->find('tr') as $element) {
$colPos = 0;
foreach ($element->find('th') as $cell) {
if (strpos(trim($cell->class), 'actions') === false && strpos(trim($cell->class), 'checker') === false) {
parseCell($cell,$mdTable,$rowPos,$colPos);
}
$colPos++;
}
foreach ($element->find('td') as $cell) {
if (strpos(trim($cell->class), 'actions') === false && strpos(trim($cell->class), 'checker') === false) {
parseCell($cell,$mdTable,$rowPos,$colPos);
}
$colPos++;
}
$rowPos++;
}
foreach ($mdTable as $key => $row) {
//clean the data
array_walk($row, "cleanCell");
fputcsv($fp, $row);
}
}
function cleanCell(&$contents,$key) {
$contents = trim($contents);
//get rid of pesky  's (aka: non-breaking spaces)
$contents = trim($contents,chr(0xC2).chr(0xA0));
$contents = str_replace(" ", "", $contents);
}
function parseCell(&$cell,&$mdTable,&$rowPos,&$colPos) {
global $repeatContentIntoSpannedCells;
//if data has already been set into the cell, skip it
while (isset($mdTable[$rowPos][$colPos])) {
$colPos++;
}
$mdTable[$rowPos][$colPos] = $cell->plaintext;
if (isset($cell->rowspan)) {
for ($i=1; $i <= ($cell->rowspan)-1; $i++) {
$mdTable[$rowPos+$i][$colPos] = ($repeatContentIntoSpannedCells ? $cell->plaintext : "");
}
}
if (isset($cell->colspan)) {
for ($i=1; $i <= ($cell->colspan)-1; $i++) {
$colPos++;
$mdTable[$rowPos][$colPos] = ($repeatContentIntoSpannedCells ? $cell->plaintext : "");
}
}
}
?>