在JPA SecondaryTable中以N：1关系保存相同的列

Question

我的代码看起来像这样：

@Data
@Accessors(chain = true, fluent = true)
@Entity
@Table(name = "T_FILE_META", indexes = {@Index(columnList = "filename")}
)
@SecondaryTable(name = "T_FILE_META_SHA",
        pkJoinColumns = @PrimaryKeyJoinColumn(name = "sha1", referencedColumnName = "sha1")
        , indexes = {@Index(columnList = "sha1", unique = true), @Index(columnList = "md5", unique = true)}
)
public class FileMeta
{
    @Id
    @GeneratedValue
    private Long id;

    private String filename;

    @Column(length = 40, table = "T_FILE_META_SHA")
    private String type;
    @Column(length = 40, table = "T_FILE_META_SHA")
    private String sha1;
    @Column(length = 32, table = "T_FILE_META_SHA")
    private String md5;
}

我想要两个表：T_FILE_META，T_FILE_META_SHA，关系是N:1

T_FILE_META( pk(id),filename, fk(sha1))
T_FILE_META_SHA( pk(sha1), md5, type)

使用此代码将获得Unable to find column with logical name: sha1 in org.hibernate.mapping.Table(t_file_meta) and its related supertables and secondary tables

关于SecondaryTable我从wiki:Tables#Multiple_tables学习。

如何定义这个实体？ BTW，T_FILE_META_SHA只有级联update。

EDITED

所以，SecondaryTable仅用于1：1，我将实体更改为

/**
 * Different filename can reference to a same file meta
 */
@Data
@Accessors(chain = true, fluent = true)
@Entity
@Table(name = "T_FILE_REF",
        indexes = {@Index(columnList = "filename"), @Index(columnList = "owner")}
)
public class FileReference
{
    @Id
    @GeneratedValue
    private Long id;

    private String filename;
    private String owner;

    /**
     * The sha1 reference may change
     * <p/>
     * This field only used for FK
     */
    @Setter(AccessLevel.NONE)
    @Getter(AccessLevel.NONE)
    private String sha1;
    @Column(columnDefinition = "default now()")
    private Date lastModificationDate;
    @Column(columnDefinition = "default now()")
    private Date uploadDate;


    @Getter(AccessLevel.NONE)
    @ManyToOne(cascade = CascadeType.PERSIST, optional = false)
    @JoinColumn(name = "sha1",
            referencedColumnName = "sha1",
            updatable = false,
            nullable = false)
    private FileMeta meta;


    public FileMeta meta()
    {
        if (meta == null)
        {
            meta = new FileMeta();
        }
        return meta;
    }

    public String md5()
    {
        return meta().md5();
    }

    public String type()
    {
        return meta().type();
    }

    public String sha1()
    {
        return meta().sha1();
    }


    @Data
    @Accessors(chain = true, fluent = true)
    @Entity
    @Table(name = "T_FILE_META", indexes = {@Index(columnList = "md5")})
    public static class FileMeta
    {
        @Id
        @Column(length = 40,
                unique = true,
                updatable = false,
                nullable = false)
        private String sha1;

        @Column(length = 32, updatable = false, nullable = false)
        private String md5;

        /**
         * File size
         */
        @Column(precision = 32, nullable = false)
        private Long length;

        /**
         * Mime type
         */
        @Column(length = 40)
        private String type;
    }
}

这看起来更明智，REF：META是N:1.REF代表META的某些领域。但这样可以吗？ FileMeta不可更新。

当我更改Reference的sha1时，是否可以自动加载FileMeta（如果存在）？

Answer 1

@SecondaryTable允许您以一对一的关系将两个表连接在一起，就像您将同一行拆分为两个单独的表一样。

对于一对多关系，您需要使用@OneToMany关联。

您的映射问题是您指定了T_FILE_META表中不存在的连接列：

pkJoinColumns = @PrimaryKeyJoinColumn(name = "sha1", referencedColumnName = "sha1")

• name：引用T_FILE_META列（不存在）
• referencedColumnName：引用T_FILE_META_SHA列

因此，您需要将两个sha1字段添加到FileMeta实体：

@Column(name = "sha1")
private String sha1Fk;

@Column(name = "sha1", table = "T_FILE_META_SHA")
private String sha1Pk;