使用Gson,我试图强制我的json对象拥有一个元素的数组值。虽然我以这种方式构造json,但是JsonParser创建了一个新的JsonObject来代替我的长度为1的数组。我系统中的大多数json都有值大于长度为1的数组并正确解析。
我在尝试使用实用程序方法强制JsonObject到JsonArray时发现了这一点。它没有像我期望的那样发挥作用。
static JsonArray bruteForceJsonArray(JsonObject object, String key) {
if(object.get(key).isJsonArray()) {
return object.get(key).getAsJsonArray();
} else {
System.out.println("broken: " + object.toString());
JsonObject temp = object.get(key).getAsJsonObject();
String fixed = "{'" + key + "':[" + temp.toString() + "]}";
System.out.println("pre-fix: " + fixed);
//parsing here reduces the array to an object - hot to prevent?
temp = new JsonParser().parse(fixed).getAsJsonObject();
System.out.println("fixed: " + object.toString());
return object.get(key).getAsJsonArray();
}
}
输出:
broken: {"token":{"NER":"O","id":1.0,"word":".","CharacterOffsetEnd":765.0,"Speaker":"PER0","POS":".","lemma":".","CharacterOffsetBegin":764.0}}
pre-fix: {'token':[{"NER":"O","id":1.0,"word":".","CharacterOffsetEnd":765.0,"Speaker":"PER0","POS":".","lemma":".","CharacterOffsetBegin":764.0}]}
fixed: {"token":{"NER":"O","id":1.0,"word":".","CharacterOffsetEnd":765.0,"Speaker":"PER0","POS":".","lemma":".","CharacterOffsetBegin":764.0}}
大多数json的一个例子,因为它们在系统中正确显示:
{"token":[{"Speaker":"PER0","POS":"IN","lemma":"except","CharacterOffsetBegin":742.0,"NER":"O","id":1.0,"word":"Except","CharacterOffsetEnd":748.0},{"CharacterOffsetEnd":752.0,"Speaker":"PER0","POS":"IN","lemma":"for","CharacterOffsetBegin":749.0,"NER":"O","id":2.0,"word":"for"},{"id":3.0,"word":"extra","CharacterOffsetEnd":758.0,"Speaker":"PER0","POS":"JJ","lemma":"extra","CharacterOffsetBegin":753.0,"NER":"O"},{"Speaker":"PER0","POS":"NN","lemma":"meat","CharacterOffsetBegin":759.0,"NER":"O","id":4.0,"word":"meat","CharacterOffsetEnd":763.0},{"CharacterOffsetEnd":764.0,"Speaker":"PER0","POS":".","lemma":".","CharacterOffsetBegin":763.0,"NER":"O","id":5.0,"word":"."}]}
有没有一个聪明的方法解决这个问题 - 某种方式我可以有效地强制我的对象进入数组,所以我可以在流中处理它们而不需要太多的try / catch或if / else处理?