recursion None python

Question

This is an XML parser code snippet which returns with None value. This is a large XML file that has a lot of subfields like this:

<root>
    <field name ="1">
        <field name ="2" showname ="ZZZ">
            <field name ="3" showname="YYY">
                <field name ="4" showname="XXX"/>
            </field>
        </field>
    </field>

The findall() finds all the elements with a tag which are direct children of the current element. I tried this, but it returned with None. It prints nothing also.

def findXXX(field):
    if field.get('name') == 'XXX' :
        return field.get('showname')
    else:
        for fieldchild in field.findall('field'):
            return findXXX(fieldchild)

If I write like this, it prints the correct value, however it returns with None.

def findXXX(field):
    if field.get('name') == 'XXX' :
        print field.get('showname')
        return field.get('showname')
    else:
        for fieldchild in field.findall('field'):
            findXXX(fieldchild)

In your first code block, I suspect having an unconditional `return` inside a `for` loop is a logic error; it will only iterate once before terminating. — Kevin, Aug 05 '14 at 12:40
Your recursion only explores the leftmost branch of the tree because you `return` immediately in your `for` loop. Either accumulate in a list and return at the end of your loop or rewrite it into a generator. — roippi, Aug 05 '14 at 12:42
BTW, do you mean to test for the `'name'` attribute or the `'showname'` attribute? You are testing against `'XXX'`. — Martijn Pieters, Aug 05 '14 at 12:48
Does this answer your question? [Why does my recursive function return None?](https://stackoverflow.com/questions/17778372/why-does-my-recursive-function-return-none) — VLAZ, Nov 03 '22 at 17:10

Martijn Pieters · Accepted Answer · 2014-08-05T12:57:59.380

In your second version, you are ignoring the return value of the recursive call. Your first version at least still returned that return value:

for fieldchild in field.findall('field'):
    return findXXX(fieldchild)

The return statement in your function returns to the direct caller, not to the caller of the outermost first invocation of findXXX().

However now only the result for the first field in the field.findall() list will return something. You probably want to return the first that is not None:

for fieldchild in field.findall('field'):
    result = findXXX(fieldchild)
    if result is not None:
        return result

This kicks of a recursive search for each field element, and returns the result of the first such recursive calls that is not None.

Note that you don't have to recurse; you can also use an XPath expression to search through the tree without having to recurse yourself:

field.find(".//field[@showname='XXX']")

This returns the first field tag whose name attribute is set to XXX, searching recursively.

Demo:

>>> from xml.etree import ElementTree as ET
>>> tree = ET.fromstring('''\
... <root>
...     <field name ="1">
...         <field name ="2" showname ="ZZZ">
...             <field name ="3" showname="YYY">
...                 <field name ="4" showname="XXX"/>
...             </field>
...         </field>
...     </field>
... </root>
... ''')
>>> tree.find(".//field[@showname='XXX']")
<Element 'field' at 0x108519bd0>
>>> tree.find(".//field[@showname='XXX']").get('name')
'4'

@user3816048: no, my answer is not exactly what you wrote. – Martijn Pieters Aug 05 '14 at 12:43 — Martijn Pieters, Aug 05 '14 at 12:43

score 0 · Answer 2 · answered Aug 05 '14 at 12:48

0

May I suggest that you use BeautifulSoup. It is easy to learn, has good documentation. http://www.crummy.com/software/BeautifulSoup/bs4/doc/

A soup object can be instantiated using the code below, where "markup" is a string containing your XML, or input file handle to the XML file:

BeautifulSoup(markup, "xml")

answered Aug 05 '14 at 12:48

coder.in.me

1,048
9
19

1

The issue is how to handle a recursion here. As it stands, this is more of a comment. – Martijn Pieters Aug 05 '14 at 12:49
My point why write XML parsing code when a well-tested module already does it. – coder.in.me Aug 05 '14 at 12:51
ElementTree and BeautifulSoup are two different approaches to providing the user with a *tree model*. – Martijn Pieters Aug 05 '14 at 12:52
But note that BeautifulSoup is better suited for HTML, really. The XML support is lovely, but still has some bugs and spotty coverage for more.. esoteric elements of the XML standard, such as processing instructions. – Martijn Pieters Aug 05 '14 at 12:56

recursion None python

2 Answers2