Not a member of Pastebin yet?
Sign Up,
it unlocks many cool features!
- <?php
- header('Content-Type: application/excel');
- header('Content-Disposition: attachment; filename="Mehedinti.csv"');
- include_once 'simple_html_dom.php';
- include_once 'csv.php';
- $urls = scrape_main_page();
- function scraping($url) {
- // create HTML DOM
- $html = file_get_html($url);
- // get article block
- if ($html && is_object($html) && isset($html->nodes)) {
- foreach ($html->find('/html/body/table') as $article) {
- // get title
- $item['titlu'] = trim($article->find('/tbody/tr[1]/td/div', 0)->plaintext);
- // get body
- $item['tr2'] = trim($article->find('/tbody/tr[2]/td[2]', 0)->plaintext);
- $item['tr3'] = trim($article->find('/tbody/tr[3]/td[2]', 0)->plaintext);
- $item['tr4'] = trim($article->find('/tbody/tr[4]/td[2]', 0)->plaintext);
- $item['tr5'] = trim($article->find('/tbody/tr[5]/td[2]', 0)->plaintext);
- $item['tr6'] = trim($article->find('/tbody/tr[6]/td[2]', 0)->plaintext);
- $item['tr7'] = trim($article->find('/tbody/tr[7]/td[2]', 0)->plaintext);
- $item['tr8'] = trim($article->find('/tbody/tr[8]/td[2]', 0)->plaintext);
- $item['tr9'] = trim($article->find('/tbody/tr[9]/td[2]', 0)->plaintext);
- $item['tr10'] = trim($article->find('/tbody/tr[10]/td[2]', 0)->plaintext);
- $item['tr11'] = trim($article->find('/tbody/tr[11]/td[2]', 0)->plaintext);
- $item['tr12'] = trim($article->find('/tbody/tr[12]/td/div/]', 0)->plaintext);
- $ret[] = $item;
- }
- // clean up memory
- $html->clear();
- unset($html);
- return $ret;}
- }
- $output = fopen("php://output", "w");
- foreach ($urls as $url) {
- $ret = scraping($url);
- foreach($ret as $v){
- fputcsv($output, $v);}
- }
- fclose($output);
- exit();
- ?>
Advertisement
Add Comment
Please, Sign In to add comment