using regex with xml2?

Question

Consider this simple example

example_xml <- '<?xml version="1.0" encoding="UTF-8"?>
<file>
<book>
<text>abracadabra</text>
<node></node>
</book>
<book>
<text>hello world</text>
<node></node>
</book>
</file>'

myxml <- xml2::read_xml(example_xml)

Now, running this works as expected

> myxml %>% xml_find_all('//book')
{xml_nodeset (2)}
[1] <book>\n  <text>abracadabra</text>\n  <node/>\n</book>
[2] <book>\n  <text>hello world</text>\n  <node/>\n</book>

but looking for nodes whose text attribute contain wor does not.

> myxml %>% xml_find_all('//book[contains(@text, "wor")]')
{xml_nodeset (0)}

What is the problem here? How can I use regex (or partial string matching) with xml2 ?

Thanks!

You do not have text attributes. You have `text` nodes inside `book` nodes. Try `xml_find_all('//book/text[contains(., "wor")]')` — Wiktor Stribiżew, Feb 07 '18 at 21:48
whoaa thanks but what is this crazy syntax `[contains(.` with a dot? — ℕʘʘḆḽḘ, Feb 07 '18 at 21:51
See [this answer](https://stackoverflow.com/a/41862373/3832970) to learn more about the difference between `text()` and `.`. BTW, maybe you need `xml_find_all(myxml, '//book[contains(., "wor")]')`? The result is `[1] \n hello world\n \n`, while the top comment solution yields `[1] hello world` — Wiktor Stribiżew, Feb 07 '18 at 21:54
I would post that as an answer. that is interesting thanks again! — ℕʘʘḆḽḘ, Feb 07 '18 at 22:01

Wiktor Stribiżew · Accepted Answer · 2018-02-07T22:19:15.543

The //book[contains(@text, "wor")] XPath finds book nodes that contain a text attribute (@ specifies an attribute) that contain wor in their values.

Your XML does not contain elements like <book text="Hello world">Title</book>, thus there are no results.

You may get the book nodes that contain wor in their text nodes using

> xml_find_all(myxml, '//book[contains(., "wor")]')
{xml_nodeset (1)}
[1] <book>\n  <text>hello world</text>\n  <node/>\n</book>

If you are fine with just text nodes as the return values, you may use

> xml_find_all(myxml, '//book/text[contains(., "wor")]')
{xml_nodeset (1)}
[1] <text>hello world</text>

If you need to get all book parents that contain any child nodes with wor text inside, use

> xml_find_all(myxml, '//*[contains(., "wor")]/parent::book')
{xml_nodeset (1)}
[1] <book>\n  <text>hello world</text>\n  <node/>\n</book>

See this answer to learn more about the difference between text() and .. In short, [contains(., "wor")] returns true if the string value of an element contains wor.

amazing. just a follow up question. How can I get all the parents nodes of the nodes that contain `wor`? — ℕʘʘḆḽḘ, Feb 07 '18 at 22:04
@ℕʘʘḆḽḘ Do you mean `xml_find_all(myxml, '//*[contains(., "wor")]/parent::book')`? See my edit with a test. — Wiktor Stribiżew, Feb 07 '18 at 22:17

using regex with xml2?

1 Answers1