jsoup for unstructured html page with table

Question

i'm trying to get the main img from this url , here what i tried so far :

Document doc = null;
    try {
        doc = Jsoup.connect(url).get();
    } catch (IOException e) {
        e.printStackTrace();
    }

    Element table = doc.select("center").get(1);
    Elements rows = table.select("table[width=970]");
    for (int i = 0; i < rows.size(); i++) {
        Element row = rows.get(1);
        Elements cols = row.select("table[width=634]");
        for (int j = 0; j < cols.size(); j++) {
            Element row1 = rows.get(1);
            Elements cols1 = row1.select("table[width=600]");
            for (int k = 0; k < cols1.size(); k++){
                Element row0 = rows.first();
                Elements cols0 = row0.select("td");
                for (Element image : cols0) {
                    String image2 = image.absUrl("src").toString();
                    Log.i("tanja7 ", "pic  " + image2);
                }
            }
        }
    }

this is the unstructured html page (i don't know how to copy the html code) What i'm doing wrong?

Answer 1

It seems that you are expecting the inner elements as result of a JSoup select method call. That is not right - you get the elements that match the selector within the "search scope", which is given by the Element(s)/document class instance from which you call select . So, if you want to get all table elements of the document you do doc.select("table") . This gives you not the rows, but the tables. Maybe you did understand this before, but your variable naming suggests otherwise.

Anyway, here is a selector that works. It will get all img elements that are (not necessarily direct) children of a table that has the attribute width=600 and is within another table of the document.

Elements imgEls = doc.select("table table[width=600] img");
System.out.println(imgEls.first().absUrl("src"));

You say the html is not structured, so you might want to check if the relevant images are really always inside two tables as specified.

update: if you are using a mobile device make sure to add:

doc = Jsoup.connect(url).userAgent("Mozilla").get();