EclipsLink 2.0的性能瓶颈

Question

最近，当我使用EclipsLink 2.0时，我遇到了持久化对象实现的性能瓶颈问题。

更具体地说，我曾经有过以下实现：

@Entity
@Table(name = "CUSTOMERS")
public class CustomerEntity implements Serializable {

@Id
@GeneratedValue(strategy = GenerationType.AUTO)
private volatile Long id;

@Column(nullable = false, unique = true)
private String name;

private static final long  serialVersionUID = 6952530957072210017L;

private String custGroup;
private String address;
private String nameOfFirstPerson;
private String contactPerson;
private String phone;
private String fax;
private String email;
private String comments;
private String defaultCustomer;
private volatile boolean delayedPaymentAllowed;
private volatile long periodOfDelayedPaymentAllowed;
private volatile boolean restrictionsOnDelayedPayment;
private volatile double maxAmoutPemittedSom;
private volatile double maxAmoutPemittedYE;
private transient String salesPointName;

@Column(length=25483)
private HashMap<String, PriceItem> totalBalance;

@Column(length=25483)
private HashMap<String, PriceItem> totalBalanceUsd;

private transient boolean valueChanged = false;


@OneToMany(mappedBy = "supplier")
private Collection<PurchaseInvoiceEntity> purchaseInvoices;

@OneToMany(mappedBy = "receiver")
private Collection<SalesInvoiceEntity> salesInvoices;

@OneToMany(mappedBy = "payer")
private Collection<PayInSlipEntity> payInSlips;

@OneToMany(mappedBy = "recipient")
private Collection<PaymentOrderEntity> paymentOrders;

@OneToMany(mappedBy = "recipient")
private Collection<WriteOffEntity> writeOffs;

@ManyToOne()
private ResponsiblePersonForDebtEntity responsiblePersonForDebt;

@ManyToOne
private CustomerGroupEntity customerGroup;


public CustomerEntity() {

    valueChanged = false;
}
...
}

当我每次将新文档的实例添加到适当的Collection中，同时将文档的新实例插入到表中时，我检测到插入文档花费的时间太长。我在使用netbeans ide 6.9的profiler模块时遇到了这个问题。实际上，我使用这些集合是为了检查相关文档是否为空。

Answer 1

为了解决这个问题，我简单地应用了以下解决方案：

1. I只是删除了文档引用：

@OneToMany(mappedBy = "supplier")私有收藏purchaseInvoices；

@OneToMany(mappedBy = "receiver")私有集合salesInvoices；

@OneToMany(mappedBy = "payer")私有收藏payInSlips；

@OneToMany(mappedBy = "recipient")私有集合paymentOrders；

@OneToMany(mappedBy =“writeOffs”)私有收藏writeOffs;

在CusotmerEntity中：

 @Entity

@Table(name = "CUSTOMERS")公共类CustomerEntity实现了可序列化{

@Id
@GeneratedValue(strategy = GenerationType.AUTO)
private volatile Long id;
@Column(nullable = false, unique = true)
private String name;
private static final long serialVersionUID = 6952530957072210017L;
private String custGroup;
private String address;
private String nameOfFirstPerson;
private String contactPerson;
private String phone;
private String fax;
private String email;
private String comments;
private String defaultCustomer;
private volatile boolean delayedPaymentAllowed;
private volatile long periodOfDelayedPaymentAllowed;
private volatile boolean restrictionsOnDelayedPayment;
private volatile double maxAmoutPemittedSom;
private volatile double maxAmoutPemittedYE;
private transient String salesPointName;
@Column(length = 25483)
private HashMap<String, PriceItem> totalBalance;
@Column(length = 25483)
private HashMap<String, PriceItem> totalBalanceUsd;
private transient boolean valueChanged = false;
@ManyToOne()
private ResponsiblePersonForDebtEntity responsiblePersonForDebt;
@ManyToOne
private CustomerGroupEntity customerGroup;

public CustomerEntity() {

    valueChanged = false;
}
 .....}

在编辑或删除CustomerEntity.时使用JPA query检查文档引用时使用

这一简单的更改消除了影响最大的性能问题。

附注：

请在使用性能分析器时，也包括对包方法的检查(例如eclipslink)。

Answer 2

有关JPA中的性能和可伸缩性问题，请阅读或收听戈登·约克的S"Strategies and Best Practices For Highly Scalable Java Persistence Applications“。