如何用R语言html_nodes匹配类名含desc-full的div元素？

Hey there! I see you're trying to pull content from elements where the class name shifts but always includes "desc-full" — let's get that sorted out for you.

Your initial attempt with .desc-full* throws an error because that's not how CSS selectors work. The * in that context isn't a wildcard for class names; instead, we need to use CSS attribute selectors or XPath to target elements whose class contains the substring we care about.

Solution 1: CSS Attribute Contains Selector #

The most straightforward fix is using a CSS selector that matches elements where the class attribute contains the string "desc-full":

# Extract elements with class containing "desc-full"
full_desc <- html_nodes(webpage, '[class*="desc-full"]')
# Convert extracted nodes to plain text
full_desc_text <- as.character(html_text(full_desc))

This selector [class*="desc-full"] will catch any element (div, p, etc.) where the class attribute has "desc-full" anywhere in its value — perfect for cases like desc-full, desc-full-v2, product-desc-full, etc.

If you want tighter precision:

• Use [class^="desc-full"] to only match classes that start with "desc-full"
• Use [class$="desc-full"] to only match classes that end with "desc-full"

Solution 2: XPath Selector #

If you prefer using XPath (fully supported by html_nodes), you can specifically target div elements with a class containing "desc-full" like this:

full_desc <- html_nodes(webpage, xpath = '//div[contains(@class, "desc-full")]')
full_desc_text <- as.character(html_text(full_desc))

This is ideal if you want to restrict your selection to only div elements, avoiding other tags that might accidentally have the same class substring.

Why Your Original Code Failed #

The .desc-full* syntax is invalid CSS — the * here would be interpreted as a universal selector (matching any element) that is a descendant of an element with class desc-full, which isn't what you intended. That's exactly why you got the Expected selector, got <DELIM '*' at 11> error.

内容的提问来源于stack exchange，提问作者Mark P.