Splinter find_by_xpath: using regex for element text()

Question

I am running a browser test with splinter and have a page with a large table. I want to locate all <tr> elements that contain a <td> with some nicely-formatted date in their text, like the one highlighted here:

It's easy to find the rows with specific text, e.g., via:

browser.find_by_xpath('//tr[.//td[contains(text(), "September")]]')

So then I tried something like the suggestions here to find text() with the general date pattern (help with simplifying my regex is welcome, too):

exp = '[A-Z][a-z]+\\s[1-9]{1,2},\\s[0-9]{4}'
browser.find_by_xpath('//tr[.//td[matches(text(), "{0}")]]'.format(exp))

This doesn't work (and I did verify that the regex works in isolation). Nor does:

browser.find_by_xpath('//tr[.//td[matches(., "{0}")]]'.format(exp))

Provided my browser allows XPath 2.0, how can I find the elements correctly?

Answer 1

Both latest Firefox and Chrome don't support XPath 2.0. Here are the relevant open issues:

• Support XPath 2.0 in XPath evaluate
• Bug 396966 - Xpath 2.0

You have to approach it without using matches(). For instance, you may filter the list of tr you find in Python by taking the EAFP approach and using datetime.strptime(). Sample:

from datetime import datetime

for tr in browser.find_by_tag("tr"):
    sorted_on = tr.find_by_css(".sorted-on")

    try:
        datetime.strptime(sorted_on.text, "%B %d, %Y")
    except ValueError:
        continue