Reads a JSON (RFC 4627) encoded value as a stream of tokens. This stream includes both literal values (strings, numbers, booleans, and nulls) as well as the begin and end delimiters of objects and arrays. The tokens are traversed in depth-first order, the same order that they appear in the JSON document. Within JSON objects, name/value pairs are represented by a single token.
Parsing JSON
To create a recursive descent parser for your own JSON streams, first create an entry point method that creates a
.
Next, create handler methods for each structure in your JSON text. You'll need a method for each object type and for each array type.
-
Within array handling methods, first call beginArray to consume the array's opening bracket. Then create a while loop that accumulates values, terminating when hasNext is false. Finally, read the array's closing bracket by calling endArray.
-
Within object handling methods, first call beginObject to consume the object's opening brace. Then create a while loop that assigns values to local variables based on their name. This loop should terminate when hasNext is false. Finally, read the object's closing brace by calling endObject.
When a nested object or array is encountered, delegate to the corresponding handler method.
When an unknown name is encountered, strict parsers should fail with an exception. Lenient parsers should call skipValue() to recursively skip the value's nested tokens, which may otherwise conflict.
If a value may be null, you should first check using peek(). Null literals can be consumed using either nextNull() or skipValue().
Example
Suppose we'd like to parse a stream of messages such as the following:
[
{
"id": 912345678901,
"text": "How do I read a JSON stream in Java?",
"geo": null,
"user": {
"name": "json_newb",
"followers_count": 41
}
},
{
"id": 912345678902,
"text": "@json_newb just use JsonReader!",
"geo": [50.454722, -104.606667],
"user": {
"name": "jesse",
"followers_count": 2
}
}
]
This code implements the parser for the above structure:
public List<Message> readJsonStream(InputStream
in)
throws IOException {
try {
return readMessagesArray(reader);
} finally {
reader.close();
}
}
public List<Message> readMessagesArray(
JsonReader reader)
throws IOException {
List<Message> messages = new ArrayList<Message>();
reader.beginArray();
while (reader.hasNext()) {
messages.add(readMessage(reader));
}
reader.endArray();
return messages;
}
public Message readMessage(
JsonReader reader)
throws IOException {
long id = -1;
String text = null;
User user = null;
List<Double> geo = null;
reader.beginObject();
while (reader.hasNext()) {
String
name = reader.nextName();
if (name.equals("id")) {
id = reader.nextLong();
} else if (name.equals("text")) {
text = reader.nextString();
} else if (name.equals("geo") && reader.peek() != JsonToken.NULL) {
geo = readDoublesArray(reader);
} else if (name.equals("user")) {
user = readUser(reader);
} else {
reader.skipValue();
}
}
reader.endObject();
return new Message(id, text, user, geo);
}
public List<Double> readDoublesArray(
JsonReader reader)
throws IOException {
List<Double> doubles = new ArrayList<Double>();
reader.beginArray();
while (reader.hasNext()) {
doubles.add(reader.nextDouble());
}
reader.endArray();
return doubles;
}
public User readUser(
JsonReader reader)
throws IOException {
String username = null;
int followersCount = -1;
reader.beginObject();
while (reader.hasNext()) {
String name = reader.nextName();
if (name.equals("name")) {
username = reader.nextString();
} else if (name.equals("followers_count")) {
followersCount = reader.nextInt();
} else {
reader.skipValue();
}
}
reader.endObject();
return new User(username, followersCount);
}
Number Handling
This reader permits numeric values to be read as strings and string values to be read as numbers. For example, both elements of the JSON array
may be read using either nextInt or nextString. This behavior is intended to prevent lossy numeric conversions: double is JavaScript's only numeric type and very large values like
cannot be represented exactly on that platform. To minimize precision loss, extremely large values should be written and read as strings in JSON.
Non-Execute Prefix
Web servers that serve private data using JSON may be vulnerable to Cross-site request forgery attacks. In such an attack, a malicious site gains access to a private JSON file by executing it with an HTML
tag.
Prefixing JSON files with ")]}'\n" makes them non-executable by
tags, disarming the attack. Since the prefix is malformed JSON, strict parsing fails when it is encountered. This class permits the non-execute prefix when lenient parsing is enabled.
Each
may be used to read a single JSON stream. Instances of this class are not thread safe.
- 作者
- Jesse Wilson
- 自从
- 1.6
1.8.8