我建议修改主构造函数，并为每个形参添加默认值:

data class Activity(
    var updated_on: String = "",
    var tags: List<String> = emptyList(),
    var description: String = "",
    var user_id: List<Int> = emptyList(),
    var status_id: Int = -1,
    var title: String = "",
    var created_at: String = "",
    var data: HashMap<*, *> = hashMapOf<Any, Any>(),
    var id: Int = -1,
    var counts: LinkedTreeMap<*, *> = LinkedTreeMap<Any, Any>()
)

您还可以通过添加?然后你可以调用null:

data class Activity(
    var updated_on: String? = null,
    var tags: List<String>? = null,
    var description: String? = null,
    var user_id: List<Int>? = null,
    var status_id: Int? = null,
    var title: String? = null,
    var created_at: String? = null,
    var data: HashMap<*, *>? = null,
    var id: Int? = null,
    var counts: LinkedTreeMap<*, *>? = null
)

一般来说，避免可空对象是一个很好的实践——以我们不需要使用它们的方式编写代码。与Java相比，非空对象是Kotlin的优点之一。因此，上面的第一个选项更可取。

这两种选择都会给你想要的结果:

val activity = Activity()
activity.title = "New Computer"
sendToServer(activity)

如果您为所有字段提供默认值，Kotlin会自动生成空构造函数。

data class User(var id: Long = -1,
                var uniqueIdentifier: String? = null)

你可以简单地调用:

val user = User()

如果你给每个主构造函数参数一个默认值:

data class Item(var id: String = "",
            var title: String = "",
            var condition: String = "",
            var price: String = "",
            var categoryId: String = "",
            var make: String = "",
            var model: String = "",
            var year: String = "",
            var bodyStyle: String = "",
            var detail: String = "",
            var latitude: Double = 0.0,
            var longitude: Double = 0.0,
            var listImages: List<String> = emptyList(),
            var idSeller: String = "")

从课堂上的实例中你可以不带参数地调用它或者用你当时的参数

var newItem = Item()

var newItem2 = Item(title = "exampleTitle",
            condition = "exampleCondition",
            price = "examplePrice",
            categoryId = "exampleCategoryId")

来自文档

注意:在JVM上，如果主构造函数的所有参数 有默认值，编译器会生成一个额外的 将使用默认值的无参数构造函数。这 使Kotlin更容易与Jackson或JPA等库一起使用 通过无参数构造函数创建类实例。

除了@miensol的答案，让我添加一些细节:

如果希望使用数据类获得java可见的空构造函数，则需要显式地定义它。

使用默认值+构造函数说明符非常简单:

data class Activity(
    var updated_on: String = "",
    var tags: List<String> = emptyList(),
    var description: String = "",
    var user_id: List<Int> = emptyList(),
    var status_id: Int = -1,
    var title: String = "",
    var created_at: String = "",
    var data: HashMap<*, *> = hashMapOf<Any, Any>(),
    var id: Int = -1,
    var counts: LinkedTreeMap<*, *> = LinkedTreeMap<Any, Any>()
) {
    constructor() : this(title = "") // this constructor is an explicit
                                     // "empty" constructor, as seen by Java.
}

这意味着通过这个技巧，您现在可以使用标准Java序列化器(Jackson、Gson等)序列化/反序列化这个对象。

你有两个选择:

Assign a default value to each primary constructor parameter: data class Activity( var updated_on: String = "", var tags: List<String> = emptyList(), var description: String = "", var user_id: List<Int> = emptyList(), var status_id: Int = -1, var title: String = "", var created_at: String = "", var data: HashMap<*, *> = hashMapOf<Any, Any>(), var id: Int = -1, var counts: LinkedTreeMap<*, *> = LinkedTreeMap<Any, Any>() ) Declare a secondary constructor that has no parameters: data class Activity( var updated_on: String, var tags: List<String>, var description: String, var user_id: List<Int>, var status_id: Int, var title: String, var created_at: String, var data: HashMap<*, *>, var id: Int, var counts: LinkedTreeMap<*, *> ) { constructor() : this("", emptyList(), "", emptyList(), -1, "", "", hashMapOf<Any, Any>(), -1, LinkedTreeMap<Any, Any>() ) }

如果你不依赖于复制或等于活动类或不使用自动生成的数据类方法，你可以使用常规类，如下所示:

class ActivityDto {
    var updated_on: String = "",
    var tags: List<String> = emptyList(),
    var description: String = "",
    var user_id: List<Int> = emptyList(),
    var status_id: Int = -1,
    var title: String = "",
    var created_at: String = "",
    var data: HashMap<*, *> = hashMapOf<Any, Any>(),
    var id: Int = -1,
    var counts: LinkedTreeMap<*, *> = LinkedTreeMap<Any, Any>()
}

并非每个DTO都需要是数据类，反之亦然。事实上，根据我的经验，我发现数据类在涉及一些复杂业务逻辑的领域特别有用。