PDF bullets are coming as question marks while parsing with Apache Tika in java

问题

I am parsing PDF files using Apache Tika (tika-app-1.3) with this code:

InputStream input = new FileInputStream("Introduction.pdf");  
AutoDetectParser parser = new AutoDetectParser();
BodyContentHandler handler = new BodyContentHandler(100 * 1024 * 1024);
Metadata metadata = new Metadata();
parser.parse(input, handler, metadata);
System.out.println(handler.toString());

handler.toString() is displaying PDF text, but this text also contains bullets, which are showing up as a ? symbol, but I want these bullets as-is. Is there any way to get the original, as-is content using Apache Tika? Or where does encoding need to be defined when parsing?

回答1:

The problem is most likely either the encoding your program is using to write to your console not matching the encoding the console is using to process the output of your program, or the font used by your console may not have a glyph for the character. See this answer.

One way to check is to print the numeric value of the characters in your string. Then you'll know if the text contains a ? (U+003F) or not.

It may even be that the PDF contains an image for the bullet instead of a character.

来源：https://stackoverflow.com/questions/17550858/pdf-bullets-are-coming-as-question-marks-while-parsing-with-apache-tika-in-java