Thank you very much, work is done ^^

According to my understanding (perhaps wrong), find is not optimal (with a lot of html files) instead of selecting fields from parent with child nodes but I dont know how to manipulate children / child_notes and so on..

Something like that :

for each loop = child of ret
do
  url1 = loop.child[1].href
  date = loop.child[2].child[1].child[1].text +   loop.child[2].child[
+1].child[2].text + loop.child[2].child[1].child[3].text
..
end loop
[download]