这些符号会使URLDecoder与UTF-8崩溃？

Question

I'm using URLDecoder to decode a string:

import java.net.URLDecoder;
URLDecoder.decode("%u6EDA%u52A8%u8F74%u627F", StandardCharsets.UTF_8.name());

Which leads to the crash

Exception in thread "main" java.lang.IllegalArgumentException: URLDecoder: Illegal hex characters in escape (%) pattern - For input string: "u6"
    at java.net.URLDecoder.decode(URLDecoder.java:194)
    at Playground$.delayedEndpoint$Playground$1(Playground.scala:45)
    at Playground$delayedInit$body.apply(Playground.scala:10)
    at scala.Function0$class.apply$mcV$sp(Function0.scala:34)
    at scala.runtime.AbstractFunction0.apply$mcV$sp(AbstractFunction0.scala:12)
    at scala.App$$anonfun$main$1.apply(App.scala:76)
    at scala.App$$anonfun$main$1.apply(App.scala:76)
    at scala.collection.immutable.List.foreach(List.scala:392)
    at scala.collection.generic.TraversableForwarder$class.foreach(TraversableForwarder.scala:35)
    at scala.App$class.main(App.scala:76)
    at Playground$.main(Playground.scala:10)
    at Playground.main(Playground.scala)

It seems like %u6 and %u8 are not allowed in the string. I've tried to read up on what these symbols are, but I've been unsuccessful. I found the string in a dataset in a field called "page title field". So I'm suspecting they are encoded symbols, I just don't know which encoding. Does anyone know what these symbols are and which encoding I should use to successfully decode them?

Answer 1

Looks like a non-standard UTF-16-based encoding of "滚动轴承", which is Chinese for "ball bearings".

I'd suggest to just .replaceAll %u by backslashes, and then use StringEscapeUtils from Apache Commons:

import org.apache.commons.lang3.StringEscapeUtils
val unescapedJava = StringEscapeUtils.unescapeJava(str.replaceAll("%u", "\\u"))
URLDecoder.decode(unescapedJava, StandardCharsets.UTF_8.name())

This should handle both kinds of escaping:

• The normal escape sequences with % followed by digits are unaffected by the replacement and unescapeJava
• The weird %u are treated specially (replaced by \\u\u003c/code> ), and eliminated in the first step.

---

If (only if) you are absolutely certain that all code points got encoded in this way, then you can do without StringEscapeUtils :

new String(
  "%u6EDA%u52A8%u8F74%u627F"
  .replaceAll("%u", "")
  .grouped(4)
  .map(Integer.parseInt(_, 16).toChar)
  .toArray
)

which produces

res: String = 滚动轴承

but I'd advice against it, because this method will break down for inputs like "%u6EDA%u52A8%u8F74%u627Fcafebabe" that contain unescaped characters. Better use a reliable library method that handles all corner cases.

Answer 2

Your string "%u6EDA%u52A8%u8F74%u627F" is syntactically wrong as a URL-encoded string. According to the javadoc of URLDecoder.decode and Wikipedia:Percent-encoding every % must be followed two hexadecimal digits.

May be you meant to use "\滚\动\轴\承" instead. This would be a syntactically correct Java string (having 4 hexadecimal escaped Unicode characters) and is equivalent to "滚动轴承" . But it still doesn't make sense to URL-decode this string. Therefore I guess the error already occurred on the encoding side, which produced this malformed URL-encoded string in the first place.