Json converts & in a String to \u0026
Matthew Harrington
I am trying to extract text from pdf and write it into a json file. While extracting unicode characters the Json converts all & to \u0026. For example my actual String is ش. (which represents ش). It prints correctly to a .txt file, to console etc. But when I try to print this string to a Json file it shows \u0026#1588;.
I am using Java, and the code is
Gson gson = new Gson();
String json = gson.toJson(pdfDoc);Note: pdfDoc is an object, that contains all the details (position, color, font.. etc) of characters inside the input PDF document. I am using gson-2.2.1.jar.
2 Answers
That's actually a valid (but not required) encoding. Any character may be encoded using the unicode escape in JSON and any valid JSON parsing library must be able to interpret those escapes.
& is not part of the characters that need encoding (see the definition of string at json.org), but there are a few JSON libraries that are quite "aggressive" in their encoding. That's not usually a problem, unless you don't really handle the resulting JSON with a conforming JSON parser.
GsonBuilder.disableHtmlEscaping() will help you turn that feature off if you absolutely need to.
2Using following code to decode \u0026 from a unsplash.com JSON file in Swift:
extension String { func utf8DecodedString()-> String { let data = self.data(using: .utf8) let message = String(data: data!, encoding: .nonLossyASCII) ?? "" return message } func utf8EncodedString()-> String { let messageData = self.data(using: .nonLossyASCII) let text = String(data: messageData!, encoding: .utf8) ?? "" return text }
}
let jsonOriginal = #"Let\u2019s not be na\357ve \u0026 dumb!"#
print(jsonOriginal)
print("----")
let jsonDecoded = jsonOriginal.utf8DecodedString()
print(jsonDecoded)
let jsonEncoded = jsonDecoded.utf8EncodedString()
print(jsonEncoded)Curiously, encoding leaves & and will not recoding to \u0026??