我尝试使用EclipseLink优化我的JPA实现。 我已经为它添加了批处理操作。但它仍然花了很多时间做50 000次插入。使用原始SQL和JDBC完成相同的插入所需的时间大约是10倍。
为了确保批处理操作实际上正常工作,我使用Wireshark检查我的数据包并且它没有使用批量插入。
这是其中一个插入数据包:
它没有这样做:
INSERT INTO ENTITYCLASSTEST (LASTNAME, NAME) VALUES ('sfirosijfhgdoi 0', 'dsufius0'), ('sfirosijfhgdoi 0', 'dsufius0'), ('sfirosijfhgdoi 0', 'dsufius0'), ('sfirosijfhgdoi 0', 'dsufius0')... and so on
我原以为它会像上面那样做,但每个数据包插入一行而不是每个数据包多行。
这是我的实体类:
@Entity
public class EntityClassTest implements Serializable {
private static final long serialVersionUID = 1L;
@Id
@GeneratedValue(strategy = GenerationType.IDENTITY)
private Long id;
private String name;
private String lastname;
public EntityClassTest() {
}
public EntityClassTest(Long id, String name, String lastname) {
this.id = id;
this.name = name;
this.lastname = lastname;
}
public EntityClassTest(String name, String lastname) {
this.name = name;
this.lastname = lastname;
}
public Long getId() {
return id;
}
public String getName() {
return name;
}
public String getLastName() {
return lastname;
}
public void setId(Long id) {
this.id = id;
}
public void setName(String name) {
this.name = name;
}
public void setLastName(String lastname) {
this.lastname = lastname;
}
@Override
public int hashCode() {
int hash = 0;
hash += (id != null ? id.hashCode() : 0);
return hash;
}
@Override
public boolean equals(Object object) {
// TODO: Warning - this method won't work in the case the id fields are not set
if (!(object instanceof EntityClassTest)) {
return false;
}
EntityClassTest other = (EntityClassTest) object;
if ((this.id == null && other.id != null) || (this.id != null && !this.id.equals(other.id))) {
return false;
}
return true;
}
@Override
public String toString() {
return "database.EntityClassTest [id=" + id + " ]";
}
}
这是我的持久化方法,它接收List
并保留其中的所有对象。
public void insertListToTable(final String persistenceUnit, final List list) throws SQLException {
final EntityManagerFactory entityManagerFactory = Persistence.createEntityManagerFactory(persistenceUnit);
final EntityManager entityManager = entityManagerFactory.createEntityManager();
final EntityTransaction transaction = entityManager.getTransaction();
try {
final int listSize = list.size();
transaction.begin();
for (int i = 0; i<listSize; i++) { //Object object : list) {
final Object object = list.get(i);
entityManager.persist(object);
if ( i % 500 == 0 ) { //500, same as the JDBC batch size defined in the persistence.xml
//flush a batch of inserts and release memory:
entityManager.flush();
entityManager.clear();
}
}
transaction.commit();
}
catch(Exception e) {
if (transaction != null) {
transaction.rollback();
}
throw new SQLException(e.getMessage());
}
finally {
entityManager.close();
}
}
我的persistence.xml
,其中我将500设置为批处理值,文件为:
<?xml version="1.0" encoding="UTF-8"?>
<persistence version="2.1" xmlns="http://xmlns.jcp.org/xml/ns/persistence" xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance" xsi:schemaLocation="http://xmlns.jcp.org/xml/ns/persistence http://xmlns.jcp.org/xml/ns/persistence/persistence_2_1.xsd">
<persistence-unit name="ExternalServer" transaction-type="RESOURCE_LOCAL">
<provider>org.eclipse.persistence.jpa.PersistenceProvider</provider>
<!-- List of Entity classes -->
<class>model.EntityClassTest</class>
<properties>
<property name="javax.persistence.jdbc.url" value="jdbc:mysql://myServer:3306/testdb?zeroDateTimeBehavior=convertToNull"/>
<property name="javax.persistence.jdbc.user" value="testdbuser"/>
<property name="javax.persistence.jdbc.driver" value="com.mysql.jdbc.Driver"/>
<property name="javax.persistence.jdbc.password" value="myPassword"/>
<property name="javax.persistence.schema-generation.database.action" value="create"/>
<!-- Weaving -->
<property name="eclipselink.weaving" value="static"/>
<!-- SQL dialect / Database type -->
<property name="hibernate.dialect" value="org.hibernate.dialect.MySQLDialect"/>
<property name="eclipselink.target-database" value="MySQL"/>
<!-- Tell the JPA provider to, by default, create the table if it does not exist. -->
<property name="javax.persistence.schema-generation.database.action" value="create"/>
<!-- No logging (For development change the value to "FINE") -->
<property name="eclipselink.logging.level" value="OFF"/>
<!-- Enable batch writing -->
<property name="eclipselink.jdbc.batch-writing" value="JDBC"/>
<!-- Batch size -->
<property name="eclipselink.jdbc.batch-writing.size" value="500"/>
</properties>
</persistence-unit>
</persistence>
所以我的问题是,为什么不批量插入?我相信我已经完成了EclipseLink的配置,这是我在EclipseLink网站上阅读的内容以及此处的内容。
////////////////////////////编辑/////////////// ///////////
正如Chris的回答所示,我将EntityClassTest
此值@GeneratedValue(strategy = GenerationType.IDENTITY)
更改为@GeneratedValue(strategy = GenerationType.SEQUENCE)
并重新运行测试,并且数据包将像以前一样发送(就像我上面发布的图片一样)。所以它并没有解决我的问题,我害怕。
////////////////////////////编辑2 ////////////// //////////
我已将persistence.xml
文件中的日志记录级别更改为FINEST
,如下所示。
<property name="eclipselink.logging.level" value="FINEST"/>
这是生成的日志。我把它放在一个pastebin中,因为它很长。
似乎很多时候都在调用Execute query InsertObjectQuery
。
////////////////////////////编辑3 ////////////// //////////
这是我使用的每个组件的版本。
+-------------------------+------------------------------+
| Variable_name | Value |
+-------------------------+------------------------------+
| innodb_version | 5.6.12 |
| protocol_version | 10 |
| slave_type_conversions | |
| version | 5.6.12-log |
| version_comment | MySQL Community Server (GPL) |
| version_compile_machine | x86_64 |
| version_compile_os | Win64 |
+-------------------------+------------------------------+
Netbeans 8.0
EclipseLink (JPA 2.1)
mysql-connector-java-5.1.24.jar
////////////////////////////编辑4 ////////////// //////////
按照CuriousMind的回答,我已将EntityClassTest
ID注释编辑为:
@Id
@GeneratedValue(strategy = GenerationType.AUTO, generator="id-seq-gen")
@SequenceGenerator( name="id-seq-gen", sequenceName="ID_SEQ_GEN", allocationSize=500 )
private Long id;
但它没有解决我的问题,我仍然每个数据包只有一个插入(如上图所示)和EclipseLink
日志我得到的:
[EL Fine]: sql: 2014-10-19 06:44:02.608--ClientSession(824177287)--Connection(1674390738)--Thread(Thread[main,5,main])--SELECT LAST_INSERT_ID()
[EL Finest]: sequencing: 2014-10-19 06:44:02.608--UnitOfWork(1985011414)--Thread(Thread[main,5,main])--assign sequence to the object (1.251 -> database.EntityClassTest [id=null ])
[EL Finest]: query: 2014-10-19 06:44:02.608--UnitOfWork(1985011414)--Thread(Thread[main,5,main])--Execute query InsertObjectQuery(database.EntityClassTest [id=null ])
[EL Finest]: query: 2014-10-19 06:44:02.608--ClientSession(824177287)--Thread(Thread[main,5,main])--Execute query ValueReadQuery(name="ID_SEQ_GEN" sql="SELECT LAST_INSERT_ID()")
[EL Fine]: sql: 2014-10-19 06:44:02.608--ClientSession(824177287)--Connection(1674390738)--Thread(Thread[main,5,main])--INSERT INTO ENTITYCLASSTEST (LASTNAME, NAME) VALUES (?, ?)
bind => [sfirosijfhgdoi 2068, dsufius1034]
[EL Fine]: sql: 2014-10-19 06:44:02.608--ClientSession(824177287)--Connection(1674390738)--Thread(Thread[main,5,main])--SELECT LAST_INSERT_ID()
[EL Finest]: sequencing: 2014-10-19 06:44:02.608--UnitOfWork(1985011414)--Thread(Thread[main,5,main])--assign sequence to the object (1.252 -> database.EntityClassTest [id=null ])
[EL Finest]: query: 2014-10-19 06:44:02.608--UnitOfWork(1985011414)--Thread(Thread[main,5,main])--Execute query InsertObjectQuery(database.EntityClassTest [id=null ])
[EL Finest]: query: 2014-10-19 06:44:02.608--ClientSession(824177287)--Thread(Thread[main,5,main])--Execute query ValueReadQuery(name="ID_SEQ_GEN" sql="SELECT LAST_INSERT_ID()")
[EL Fine]: sql: 2014-10-19 06:44:02.608--ClientSession(824177287)--Connection(1674390738)--Thread(Thread[main,5,main])--INSERT INTO ENTITYCLASSTEST (LASTNAME, NAME) VALUES (?, ?)
bind => [sfirosijfhgdoi 2244, dsufius1122]
等等......
答案 0 :(得分:4)
您正在使用GenerationType.IDENTITY进行排序,这需要逐个检索每个插入语句中的ID。尝试一个允许批量预分配的排序方案,您将看到改进: http://en.wikibooks.org/wiki/Java_Persistence/Identity_and_Sequencing#Identity_sequencing
答案 1 :(得分:1)
已经有一年了,可能为时已晚,无法回答。在我的情况下,我发现flush()导致了问题。我在每条记录的persist()之后调用flush()。这可以防止批量写入进行优化并导致插入性能不佳。删除flush()后,一切顺利。
答案 2 :(得分:0)
似乎序列生成导致了这个问题,你可以参考这个post。它提供了预分配序列生成的方法。
答案 3 :(得分:0)
这可能是因为序列/缩进(需要使用预分配),但对于mysql我认为你需要一个连接字符串属性来允许批量插入:
rewriteBatchedStatements =真
jdbc:mysql://localhost:3306/test?rewriteBatchedStatements=true