Spring Data DynamoDB - 存储库执行扫描而不是查询

Question

预期行为

Repository 对只使用 hashKey 和 rangeKey 属性的方法执行查询，结果如下：

"{"TableName":"music","KeyConditionExpression":"artist = :_artist and begins_with(title, :_title)","ExpressionAttributeValues":{":_artist":{"S":"Djavan"},":_title":{"S":"Eu te devoro"}}}"

实际行为

Repository 对仅使用 hashKey 和 rangeKey 属性的方法执行 scanFilter，结果如下：

"{"TableName":"music","ScanFilter":{"artist":{"AttributeValueList":[{"S":"Djavan"}],"ComparisonOperator":"EQ"},"title":{"AttributeValueList":[{"S":"Eu te devoro"}],"ComparisonOperator":"BEGINS_WITH"}}}"

重现问题的步骤

使用名为 Music 的实体

@DynamoDBTable(tableName = "music")
data class Music(
        @field:Id
        @DynamoDBIgnore
        val id: MusicId = MusicId(),

        var genre: String? = null
) {
    @DynamoDBHashKey(attributeName = "artist")
    fun getArtist() = id.artist

    fun setArtist(artist: String) {
        id.artist = artist
    }

    @DynamoDBHashKey(attributeName = "title")
    fun getTitle() = id.title

    fun setTitle(title: String) {
        id.title = title
    }

}

@DynamoDBDocument
data class MusicId(
        @field:DynamoDBHashKey(attributeName = "artist")
        var artist: String? = null,

        @field:DynamoDBRangeKey(attributeName = "title")
        var title: String? = null
) : Serializable

还有一个仓库

@EnableScan //I know that if I remove this annotation, enables won't be permitted, but the problem is that the implementation code doesn't recognize my method as a key query and if I remove this annotation, the method falls on invocation
interface MusicRepository : CrudRepository<Music, MusicId> {

    fun findByArtistAndTitleStartingWith(artista: String, sortKey: String): List<Music>
}

当我调用时：

@PostConstruct
    fun init() {
        println(musicRepository.findByArtistAndTitleStartingWith("Djavan", "Eu te devoro").joinToString())
    }

如上所示，日志显示我对 AWS 的调用

规格

• 库：https://github.com/boostchicken/spring-data-dynamodb
• Spring Data DynamoDB 版本：5.2.5
• Spring 数据版本：未使用
• Spring Boot Starter Web 版本：2.3.4.RELEASE
• AWS 开发工具包版本：1.11.573
• Java 版本：11
• 平台详情：Windows

Github issue

我有什么问题吗？还是spring数据创建对aws的正确查询的另一种方法？

Answer 1

尝试随机更改实体后。我在存储库方法中得到了预期的结果。

使用复合键映射实体的正确方法

@DynamoDBTable(tableName = "music")
data class Music(

        @get:DynamoDBHashKey(attributeName = "artist")
        var artist: String? = null,

        @get:DynamoDBRangeKey(attributeName = "title")
        var title: String? = null,

        var genre: String? = null
) {

    @Id
    private var id: MusicId? = null
        get() = MusicId(artist, title)
}

@DynamoDBDocument
data class MusicId(

        @field:DynamoDBHashKey(attributeName = "artist")
        var artist: String? = null,

        @field:DynamoDBRangeKey(attributeName = "title")
        var title: String? = null
) : Serializable

更改实体并调用后：

musicRepository.findByArtistAndTitleStartingWith("Djavan", "Eu te devoro")

我得到了正确的 AWS api 调用。

"{"TableName":"music","ConsistentRead":true,"KeyConditions":{"artist":{"AttributeValueList":[{"S":"Djavan"}],"ComparisonOperator":"EQ"},"title":{"AttributeValueList":[{"S":"Eu te devoro"}],"ComparisonOperator":"BEGINS_WITH"}},"ScanIndexForward":true}"