这可以通过自定义TypeAdapterFactory来解决，它获取字段类型的默认适配器，但 Package 它并首先检查JSON数据以检测无效值，例如您的示例中的空字符串。举例来说：

class SwitchMgmtAdapterFactory implements TypeAdapterFactory {
    // Constructor called by Gson using reflection
    public SwitchMgmtAdapterFactory() {
    }

    @Override
    public <T> TypeAdapter<T> create(Gson gson, TypeToken<T> type) {
        // Obtain the default adapter for the type of the field
        // Uses getAdapter instead of getDelegateAdapter due to https://github.com/google/gson/issues/1028
        TypeAdapter<T> delegate = gson.getAdapter(type);

        return new TypeAdapter<T>() {
            @Override
            public T read(JsonReader in) throws IOException {
                if (in.peek() == JsonToken.BEGIN_OBJECT) {
                    return delegate.read(in);
                }

                // Otherwise handle malformed value
                // In this example here the value is skipped and `null` is returned,
                // but could instead also inspect the JSON value and use it in some way
                in.skipValue();
                return null;
            }

            @Override
            public void write(JsonWriter out, T value) throws IOException {
                delegate.write(out, value);
            }
        };
    }
}

然后，您可以使用@JsonAdapter在现场注册该适配器工厂：

public class Site 
{
    @JsonAdapter(SwitchMgmtAdapterFactory.class)
    @SerializedName("switch_mgmt")
    Map<String, Object> switchMgmt;

    ...
}

然而，正如在评论中提到的，我个人认为要求客户提供有效数据会更好。这使您更容易维护代码，也使客户端更容易注意到它们发送的数据格式不正确。否则，他们可能会浪费时间试图理解为什么他们提供的价值被忽略。虽然也注意到Gson目前是does not detect unknown fields和missing fields。