Java 用OPENJPA加载巨树

我必须用JPA加载一棵树（文件夹和文件）。但是有1000多个节点，加载整个树需要14秒

Categorie.java：

public class Categorie implements Serializable {
private static final long serialVersionUID = 1L;
@Id
@GeneratedValue(strategy = GenerationType.IDENTITY)
private int idcategorie;
@ManyToOne()
@JoinColumn(name = "IDCATPARENT")
private Categorie categorieParent;

@OneToMany(mappedBy = "categorieParent", cascade = ALL)
private List<Categorie> listeCategorie = new ArrayList<Categorie>();

@OneToMany(mappedBy = "categorieParent", cascade = ALL)
private List<Fichier> listeFichier = new ArrayList<Fichier>();

数据存储在DB2中，我用于加载树：

SELECT c FROM Categorie c LEFT JOIN FETCH c.listeCategorie
SELECT c FROM Categorie c LEFT JOIN FETCH c.listeFichier

我怎样才能更快地加载它

谢谢，对不起，我的英语

编辑： 由OpenJPA生成的查询

SELECT t0, t1, t2 FROM SCHEMA.Categorie t0 LEFT OUTER JOIN SCHEMA.Categorie t1 ON t0.IDCATPARENT = t1.idcategorie LEFT OUTER JOIN SCHEMA.Categorie t2 ON t0.idcategorie = t2.IDCATPARENT WHERE (t0.supprimer = CAST(? AS BIGINT)) ORDER BY t2.IDCATPARENT ASC 
openjpa.jdbc.SQL - <t 1860790064, conn 339897564> [56 ms] spent

SELECT t0, t1, t2 FROM SCHEMA.Categorie t0 LEFT OUTER JOIN SCHEMA.Categorie t1 ON t0.IDCATPARENT = t1.idcategorie LEFT OUTER JOIN SCHEMA.Fichier t2 ON t0.idcategorie = t2.IDCATEGORIE WHERE (t0.supprimer = CAST(? AS BIGINT)) ORDER BY t2.IDCATEGORIE ASC  
openjpa.jdbc.SQL - <t 1860790064, conn -1256330574> [210 ms] spent

从架构中选择t0、t1、t2。将t0分类为左外部联接架构。将t0.IDCATPARENT上的t1分类为t1.IDCATPARENT左外部联接架构。将t0.IDCATPARENT上的t2分类为t2.IDCATPARENT，其中（t0.Supplier=CAST（？AS BIGINT））按t2.IDCATPARENT ASC排序
openjpa.jdbc.SQL-[56毫秒]花费
从架构中选择t0、t1、t2。分类t0左外部联接架构。分类t1在t0.IDCATPARENT=t1.IDCATEGRIE左外部联接架构上。Fichier t2在t0.IDCATEGRIE=t2.IDCATEGRIME其中（t0.Supplier=CAST（？AS BIGINT））按t2.IDCATEGRIE ASC排序
openjpa.jdbc.SQL-[210毫秒]花费

---

因此，调用数据库非常快，但之后，需要10-15秒才能获得结果作为列表…

您是否尝试过延迟加载集合？例如，将连接注释修改为

@OneToMany（fetch=FetchType.LAZY，mappedBy=“categorieParent”，cascade=ALL）

。另外，为什么需要加载所有节点？默认情况下，OnetoMany的fetchtype是惰性的，我已经尝试切换此模式，但结果大致相同。我必须提供一个管理接口来管理树，并且必须将所有节点导出到jSon。所以我必须加载所有树：（我首先启用openjpasql日志记录（OpenJPA.Log=SQL=trace）来查看是否有任何东西跳出你的视线

SELECT t0, t1, t2 FROM SCHEMA.Categorie t0 LEFT OUTER JOIN SCHEMA.Categorie t1 ON t0.IDCATPARENT = t1.idcategorie LEFT OUTER JOIN SCHEMA.Categorie t2 ON t0.idcategorie = t2.IDCATPARENT WHERE (t0.supprimer = CAST(? AS BIGINT)) ORDER BY t2.IDCATPARENT ASC 
openjpa.jdbc.SQL - <t 1860790064, conn 339897564> [56 ms] spent

SELECT t0, t1, t2 FROM SCHEMA.Categorie t0 LEFT OUTER JOIN SCHEMA.Categorie t1 ON t0.IDCATPARENT = t1.idcategorie LEFT OUTER JOIN SCHEMA.Fichier t2 ON t0.idcategorie = t2.IDCATEGORIE WHERE (t0.supprimer = CAST(? AS BIGINT)) ORDER BY t2.IDCATEGORIE ASC  
openjpa.jdbc.SQL - <t 1860790064, conn -1256330574> [210 ms] spent