JPA and Hibernate Performance Tips

@vlad_mihalcea vladmihalcea.com
JPA and Hibernate
Performance Tips

About me
vladmihalcea.com

Mappings
CC BY-SA 2.0 - https://www.flickr.com/photos/47515486@N05/44129053595/

TABLE generator
@Entity
public class Post {
@Id
@GeneratedValue(strategy = GenerationType.TABLE)
private Long id;
private String title;
//Getters and setters omitted for brevity
}

TABLE generator
CREATE TABLE hibernate_sequences (
sequence_name (255) NOT NULL,
next_val,
PRIMARY KEY (sequence_name)
)
CREATE TABLE post (
id NOT NULL,
title (255),
PRIMARY KEY (id)
)

TABLE generator
public IntegralDataTypeHolder getNextValue() {
return session.getTransactionCoordinator().createIsolationDelegate()
.delegateWork(new AbstractReturningWork<IntegralDataTypeHolder>() {
public IntegralDataTypeHolder execute(
Connection connection) throws SQLException {
…
}, true
);
}
for (int i = 0; i < 3; i++) {
Post post = new Post();
post.setTitle(
String.format("High-Performance Java Persistence, Part %d", i + 1)
);
entityManager.persist(post);
}

SELECT tbl.next_val FROM hibernate_sequences tbl
WHERE tbl.sequence_name = 'default' FOR UPDATE
INSERT INTO hibernate_sequences (sequence_name, next_val) VALUES ('default', 1)
UPDATE hibernate_sequences SET next_val = 2
WHERE next_val = 1 AND sequence_name = 'default'
TABLE generator

INSERT INTO post (title, id)
VALUES ('High-Performance Java Persistence, Part 1', 1)
TABLE generator

IDENTITY vs TABLE generator

SEQUENCE vs TABLE generator

AUTO Generator
@Entity
public class Post {
@Id
@GeneratedValue(strategy = GenerationType.AUTO)
private Long id;
}

AUTO Generator
• Prior to Hibernate 5 – native strategy.
• Hibernate 5 – SequenceStyleGenerator strategy (falls back to
TABLE)

AUTO Generator – Hibernate 5 and MySQL
@Id
@GeneratedValue(generator="native")
@GenericGenerator(name = "native", strategy = "native")
private Long id;
INSERT INTO post (title)
VALUES ('High-Performance Java Persistence, Part 1')

Identifier portability – annotation mapping
@Entity(name = "Post")
@Table(name = "post")
public class Post {
@Id
@GeneratedValue(generator = "sequence",
strategy = GenerationType.SEQUENCE)
@SequenceGenerator(name = "sequence", allocationSize = 10)
private Long id;
}

Identifier portability – XML mapping
<?xml version="1.0" encoding="UTF-8"?>
<entity-mappings
xmlns="http://xmlns.jcp.org/xml/ns/persistence/orm"
xmlns:xsi=http://www.w3.org/2001/XMLSchema-instance
xsi:schemaLocation="http://xmlns.jcp.org/xml/ns/persistence/orm_2_2.xsd"
version="2.2">
<package>com.vladmihalcea.book.hpjp.hibernate.identifier.global</package>
<entity class="Post" access="FIELD">
<attributes>
<id name="id">
<generated-value strategy="IDENTITY"/>
</id>
</attributes>
</entity>
</entity-mappings

Custom types – IP address column
• IP address stored in the Classless Inter-Domain Routing format
• VARCHAR(18)
• BIGINT column encoding – 4 bytes (e.g. 192.168.123.231) + 1 byte (e.g. 24)
• PostgreSQL cidr or inet type (requires 7 bytes)

Custom types – PostgreSQL inet column
Event matchingEvent = (Event) entityManager
.createNativeQuery(
"SELECT e.* " +
"FROM event e " +
"WHERE " +
" e.ip && CAST(:network AS inet) = TRUE")
.setParameter("network", "192.168.0.1/24")
.getSingleResult();
assertEquals("192.168.0.123", matchingEvent.getIp().getAddress());

Custom types – PostgreSQL inet column
@Entity(name = "Event")
@Table(name = "event")
@TypeDef(typeClass = IPv4Type.class, defaultForType = IPv4.class)
public class Event {
@Id
@GeneratedValue
private Long id;
@Column(name = "ip", columnDefinition = "inet")
private IPv4 ip;
}

The hibernate-types project
<dependency>
<groupId>com.vladmihalcea</groupId>
<artifactId>hibernate-types-52</artifactId>
<version>${hibernate-types.version}</version>
</dependency>
<dependency>
<groupId>com.vladmihalcea</groupId>
<artifactId>hibernate-types-5</artifactId>
<version>${hibernate-types.version}</version>
</dependency>

JSON for unstructured data
@Entity(name = "Book")
@Table(name = "book")
@TypeDef(typeClass = JsonNodeBinaryType.class, defaultForType = JsonNode.class)
public class Book {
@Id
@GeneratedValue
private Long id;
@NaturalId
private String isbn;
@Column(columnDefinition = "jsonb")
private JsonNode properties;
}

Book book = new Book();
book.setIsbn("978-9730228236");
book.setProperties(
JacksonUtil.toJsonNode(
"{" +
" "title": "High-Performance Java Persistence"," +
" "author": "Vlad Mihalcea"," +
" "publisher": "Amazon"," +
" "price": 44.99" +
"}"
)
);

Book book = entityManager.unwrap(Session.class).bySimpleNaturalId(Book.class)
.load("978-9730228236");
book.setProperties(
JacksonUtil.toJsonNode(
"{" +
" "title": "High-Performance Java Persistence"," +
" "author": "Vlad Mihalcea"," +
" "publisher": "Amazon"," +
" "price": 44.99," +
" "url": "https://www.amzn.com/973022823X/"" +
"}"
)
);

Statement caching
CC BY 2.0 - https://www.flickr.com/photos/southbeachcars/16065861514/

Execution plan cache

Oracle server-side statement caching
• Soft parse
• Hard parse
• Bind peeking
• Adaptive cursor sharing (since 11g)

SQL Server server-side statement caching
• Execution plan cache
• Parameter sniffing
• Prepared statements should use the qualified object name
SELECT *
FROM etl.dbo.task
WHERE status = ?

PostgreSQL server-side statement caching
• Prior to 9.2 – execution plan cache
• 9.2 – deferred optimization
• The prepareThreshold connection property

MySQL server-side statement caching
• No execution plan cache
• Since Connector/J 5.0.5 PreparedStatements are emulated
• To activate server-side prepared statements:
• useServerPrepStmts
• cachePrepStmts

Client-side statement caching
• Recycling Statement, PreparedStatement or
CallableStatement objects
• Reusing database cursors

Oracle implicit client-side statement caching
• Connection-level cache
• PreparedStatement and CallabledStatement only
• Caches metadata only
connectionProperties.put(
"oracle.jdbc.implicitStatementCacheSize",
Integer.toString(cacheSize));
dataSource.setConnectionProperties(connectionProperties);

Oracle implicit client-side statement caching
• Once enabled, all statements are cached.
• Can be disabled on a per statement basis
if (statement.isPoolable()) {
statement.setPoolable(false);
}

Oracle explicit client-side statement caching
• Caches both metadata, execution state and data
OracleConnection oracleConnection =
(OracleConnection) connection;
oracleConnection.setExplicitCachingEnabled(true);
oracleConnection.setStatementCacheSize(cacheSize);

Oracle explicit client-side statement caching
PreparedStatement statement = oracleConnection.
getStatementWithKey(SELECT_POST_KEY);
if (statement == null) {
statement = connection.prepareStatement(SELECT_POST);
}
try {
statement.setInt(1, 10);
statement.execute();
} finally {
((OraclePreparedStatement) statement).closeWithKey(SELECT_POST_KEY);
}

SQL Server client-side statement caching
• The SQL Server JDBC driver version 6.3 added support for statement
caching.
• Prepared Statement caching is disabled by default.
connection.setStatementPoolingCacheSize(10);
connection.setDisableStatementPooling(false);

PostgreSQL Server client-side statement caching
• PostgreSQL JDBC Driver 9.4-1202 makes client-side statement
connection-bound instead of statement-bound
• Configurable:
• preparedStatementCacheQueries (default is 256)
• preparedStatementCacheSizeMiB (default is 5MB)
• Statement.setPoolable(false) is not supported

MySQL Server client-side statement caching
• Configurable:
• cachePrepStmts (default is false)
Required for server-side statement caching as well
• prepStmtCacheSize (default is 25)
• prepStmtCacheSqlLimit (default is 256)
• Statement.setPoolable(false) works for server-side
statements only

Statement caching gain (one minute interval)
Database System No Caching Throughput
(SPM)
Caching Throughput
(SPM)
Percentage Gain
DB_A 419 833 507 286 20.83%
DB_B 194 837 303 100 55.56%
DB_C 116 708 166 443 42.61%
DB_D 15 522 15 550 0.18%

Queries
CC BY 2.0 - https://www.flickr.com/photos/justinbaeder/5317820857/

Criteria API literal handling
CriteriaBuilder cb = entityManager.getCriteriaBuilder();
CriteriaQuery<Book> cq = cb.createQuery(Book.class);
Root<Book> root = cq.from(Book.class);
cq.select(root);
cq.where(cb.equal(root.get("name"), "High-Performance Java Persistence"));
Book book = entityManager.createQuery(cq).getSingleResult();
SELECT
b.id AS id1_0_, b.isbn AS isbn2_0_, b.name AS name3_0_
FROM
book b
WHERE
b.name = ?

CriteriaBuilder cb = entityManager.getCriteriaBuilder();
CriteriaQuery<Book> cq = cb.createQuery(Book.class);
Root<Book> root = cq.from(Book.class);
cq.select(root);
cq.where(cb.equal(root.get("isbn"), 978_9730228236L));
Book book = entityManager.createQuery(cq).getSingleResult();
SELECT
b.id AS id1_0_, b.isbn AS isbn2_0_, b.name AS name3_0_
FROM
book b
WHERE
b.isbn = 9789730228236

IN query padding optimization
List<Post> getPostByIds(EntityManager entityManager, Integer... ids) {
return entityManager.createQuery(
"select p " +
"from Post p " +
"where p.id in :ids", Post.class)
.setParameter("ids", Arrays.asList(ids))
.getResultList();
}

SELECT p.id AS id1_0_, p.title AS title2_0_
FROM post p
WHERE p.id IN (?, ?, ?)
FROM post p
WHERE p.id IN (?, ?, ?, ?)
assertEquals(3, getPostByIds(entityManager, 1, 2, 3).size());
assertEquals(4, getPostByIds(entityManager, 1, 2, 3, 4).size());

FROM post p
WHERE p.id IN (?, ?, ?, ?, ?)
FROM post p
WHERE p.id IN (?, ?, ?, ?, ?, ?)
assertEquals(5, getPostByIds(entityManager, 1, 2, 3, 4, 5).size());
assertEquals(6, getPostByIds(entityManager, 1, 2, 3, 4, 5, 6).size());

<property name="hibernate.query.in_clause_parameter_padding" value="true" />
FROM post p
WHERE p.id IN (?, ?, ?, ?) -- Params: (1, 2, 3, 3)
FROM post p
WHERE p.id IN (?, ?, ?, ?) -- Params: (1, 2, 3, 4)
assertEquals(3, getPostByIds(entityManager, 1, 2, 3).size());
assertEquals(4, getPostByIds(entityManager, 1, 2, 3, 4).size());

assertEquals(5, getPostByIds(entityManager, 1, 2, 3, 4, 5).size());
FROM post p
WHERE p.id IN (?, ?, ?, ?, ?, ?, ?, ?) -- Params: (1, 2, 3, 4, 5, 5, 5, 5)
assertEquals(6, getPostByIds(entityManager, 1, 2, 3, 4, 5, 6).size());
FROM post p
WHERE p.id IN (?, ?, ?, ?, ?, ?, ?, ?) -- Params: (1, 2, 3, 4, 5, 6, 6, 6)

Query plan cache
• Entity queries (JPQL, Criteria API) need to be compiled to SQL
• Configurable:
• hibernate.query.plan_cache_max_size (2048)
• hibernate.query.plan_parameter_metadata_max_size (128)

Entity query plan cache improvement

Native SQL query plan cache improvement

JPQL DISTINCT scalar query
List<Integer> publicationYears = entityManager.createQuery(
"select distinct year(p.createdOn) " +
"from Post p " +
"order by year(p.createdOn)", Integer.class)
.getResultList();
SELECT DISTINCT
extract(YEAR FROM p.created_on) AS col_0_0_
FROM post p
ORDER BY (YEAR FROM p.created_on)

JPQL DISTINCT entity query
List<Post> posts = entityManager.createQuery(
"select distinct p " +
"from Post p " +
"left join fetch p.comments " +
"where p.title = :title", Post.class)
.setParameter("title", "High-Performance Java Persistence")
.getResultList();
SELECT DISTINCT
p.id AS id1_0_0_, pc.id AS id1_1_1_, p.title AS title2_0_0_,
pc.review AS review2_1_1_, pc.post_id AS post_id3_1_0__,
pc.id AS id1_1_0__
FROM post p
LEFT OUTER JOIN post_comment pc ON p.id=pc.post_id
WHERE p.title = ?

"select distinct p " +
"from Post p " +
"left join fetch p.comments " +
"where p.title = :title", Post.class)
.setParameter("title", "High-Performance Java Persistence")
.setHint(QueryHints.HINT_PASS_DISTINCT_THROUGH, false)
.getResultList();
SELECT
p.id AS id1_0_0_, pc.id AS id1_1_1_, p.title AS title2_0_0_,
pc.review AS review2_1_1_, pc.post_id AS post_id3_1_0__,
pc.id AS id1_1_0__
FROM post p
LEFT OUTER JOIN post_comment pc ON p.id=pc.post_id
WHERE p.title = ?

Fetching
CC BY 2.0 - https://www.flickr.com/photos/bala_/3544603505/

Persistence Context – loaded state

Persistence Context – dirty checking

Persistence Context size
//session-level configuration
Session session = entityManager.unwrap(Session.class);
session.setDefaultReadOnly(true);
//query-level configuration
"select p from Post p", Post.class)
.setHint(QueryHints.HINT_READONLY, true)
.getResultList();

@Transactional(readOnly = true)
public List<Post> findAllByTitle(String title) {
List<Post> posts = postDAO.findByTitle(title);
org.hibernate.engine.spi.PersistenceContext persistenceContext =
getHibernatePersistenceContext();
for(Post post : posts) {
assertTrue(entityManager.contains(post));
EntityEntry entityEntry = persistenceContext.getEntry(post);
assertNull(entityEntry.getLoadedState());
}
return posts;
}
Spring 5.1 read-only optimization

Fetching – Pagination
• JPA / Hibernate API works for both entity and native SQL queries
List<PostCommentSummary> summaries =
entityManager.createQuery(
"select new PostCommentSummary( " +
" p.id, p.title, c.review ) " +
"from PostComment c " +
"join c.post p")
.setFirstResult(pageStart)
.setMaxResults(pageSize)
.getResultList();

Fetching – 100k vs 100 rows
Fetch all Fetch limit
0
500
1000
1500
2000
2500
3000
3500
4000
4500
5000
Time(ms)
DB_A DB_B DB_C DB_D

JPA 2.2 Streaming

JPA 2.2 Streaming
default Stream getResultStream() {
return getResultList().stream();
}
final ScrollableResultsImplementor scrollableResults = scroll(
ScrollMode.FORWARD_ONLY
);

JDBC-level streaming support
• You still need to consider the Statement.setFetchSize.
• On MySQL, you need to set it to Integer.MIN_VALUE.
• On PostgreSQL, you need to set it to a positive integer value.
• What about the Execution Plan?

JPA 2.2 Streaming – Execution Plans
List<String> executionPlanLines = doInJPA(entityManager -> {
try(Stream<String> postStream = entityManager
.createNativeQuery(
"EXPLAIN ANALYZE " +
"SELECT p " +
"FROM post p " +
"ORDER BY p.created_on DESC")
.setHint(QueryHints.HINT_FETCH_SIZE, 50)
.getResultStream()
) {
return postStream.limit(50).collect(Collectors.toList());
}
});
CREATE INDEX idx_post_created_on ON post (created_on DESC)

LOGGER.info( "Execution plan: {}",
executionPlanLines
.stream()
.collect(Collectors.joining( "n" ))
);
Execution plan:
Sort (cost=65.53..66.83 rows=518 width=564)
(actual time=2.876..3.399 rows=5000 loops=1)
Sort Key: created_on DESC
Sort Method: quicksort Memory: 896kB
-> Seq Scan on post p
(cost=0.00..42.18 rows=518 width=564)
Planning time: 1.586 ms
Execution time: 4.061 ms

List<String> executionPlanLines = doInJPA(entityManager -> {
return entityManager
.createNativeQuery(
"EXPLAIN ANALYZE " +
"SELECT p " +
"FROM post p " +
"ORDER BY p.created_on DESC")
.setMaxResults(50)
.getResultList();
});
LOGGER.info("Execution plan: {}",
executionPlanLines
.stream()
.collect(Collectors.joining("n"))
);

Execution plan:
Limit (cost=0.28..25.35 rows=50 width=564)
-> Index Scan using idx_post_created_on on post p
(cost=0.28..260.04 rows=518 width=564)
Planning time: 1.511 ms
Execution time: 0.148 ms

Fetching – Open Session in View Anti-Pattern

Fetching – Temporary Session Anti-Pattern
• “Band aid” for LazyInitializationException
• One temporary Session/Connection for every lazily fetched
association
<property
name="hibernate.enable_lazy_load_no_trans"
value="true"/>

Batching
CC BY-SA 2.0 - https://www.flickr.com/photos/dozodomo/6975654335/

JDBC Batch Updates

Transactional write-behind cache

Action queue

Batch processing – flush-clear-commit
try {
entityManager.getTransaction().begin();
for ( int i = 0; i < entityCount; ++i ) {
if ( i > 0 && i % batchSize == 0 ) {
flush( entityManager );
}
entityManager.persist( new Post( String.format( "Post %d", i + 1 ) ) );
}
entityManager.getTransaction().commit();
} catch (RuntimeException e) {
if ( entityManager.getTransaction().isActive()) {
entityManager.getTransaction().rollback();
}
throw e;
} finally {
entityManager.close();
}

private void flush(EntityManager entityManager) {
entityManager.flush();
entityManager.clear();
entityManager.getTransaction().commit();
entityManager.getTransaction().begin();
}
Batch processing – flush and commit

Hibernate batching – default behavior
for (int i = 0; i < 3; i++) {
entityManager.persist(
new Post(String.format("Post no. %d", i + 1))
);
}
INSERT INTO post (title, id) VALUES ('Post no. 1', 1)

Enable JDBC batching
<property name="hibernate.jdbc.batch_size" value="5"/>
Query: ["INSERT INTO post (title, id) VALUES (?, ?)"],
Params: [('Post no. 1', 1),
('Post no. 2', 2),
('Post no. 3', 3)]
entityManager.unwrap(Session.class).setJdbcBatchSize(10);
for (long i = 0; i < 10; ++i) {
Post post = new Post();
post.setTitle(String.format("Post nr %d", i));
entityManager.persist(post);
}

PostgreSQL batch statements
log_statement = 'all'
LOG: execute S_2: insert into post (title, id) values ($1, $2)
DETAIL: parameters: $1 = 'Post no. 1', $2 = '1'

PostgreSQL rewrite batch statements
PGSimpleDataSource dataSource = (PGSimpleDataSource) dataSource();
dataSource.setReWriteBatchedInserts(true);
LOG: execute <unnamed>: insert into post (title, id)
values ($1, $2),($3, $4),($5, $6)
DETAIL: parameters: $1 = 'Post no. 1', $2 = '1’,
$3 = 'Post no. 2', $4 = '2’,
$5 = 'Post no. 3', $6 = '3'

Default UPDATE - Post entity mapping
public class Post {
@Id
private Long id;
private long likes;
}

Default UPDATE - Post entity data
Post post1 = new Post();
post1.setId(1L);
post1.setTitle("High-Performance Java Persistence");
entityManager.persist(post1);
Post post2 = new Post();
post2.setId(2L);
post2.setTitle("Java Persistence with Hibernate");
entityManager.persist(post2);

Default UPDATE - statement batching
Post post1 = entityManager.find(Post.class, 1L);
post1.setTitle("High-Performance Java Persistence 2nd Edition");
post2.setLikes(12);
Query :[
"update post set likes=?, title=? where id=?"
],
Params:[
(0, High-Performance Java Persistence 2nd Edition, 1),
(12, Java Persistence with Hibernate, 2)
]

Default UPDATE disadvantages
• Column size
• Indexes
• Replication
• Undo and redo log
• Triggers

Hibernate dynamic update
@DynamicUpdate
public class Post {
@Id
private Long id;
private long likes;
}

Hibernate dynamic update
Query:["update post set title=? where id=?"],
Params:[(High-Performance Java Persistence 2nd Edition, 1)]
Query:["update post set likes=? where id=?"],
Params:[(12, 2)]
post1.setTitle("High-Performance Java Persistence 2nd Edition");
post2.setlikes(12);

Updating detached entities - Post and PostComment
List<Post> posts = doInJPA(entityManager -> {
return entityManager.createQuery(
"select p " +
"from Post p " +
"join fetch p.comments ", Post.class)
.getResultList();
});
for (Post post: posts) {
post.setTitle("Vlad Mihalcea's " + post.getTitle());
for (PostComment comment: post.getComments()) {
comment.setReview(comment.getReview() + " read!");
}
}

• JPA merge
• Hibernate update
JPA merge and Hibernate update
entityManager.merge(post);
}
Session session = entityManager.unwrap(Session.class);
session.update(post);
}

JPA merge – SELECT statements
SELECT p.id AS id1_0_1_, p.title AS title2_0_1_, c.post_id AS post_id3_1_3_,
c.id AS id1_1_3_, c.review AS review2_1_0_
FROM post p
LEFT OUTER JOIN post_comment c ON p.id = c.post_id
WHERE p.id = 1
FROM post p
WHERE p.id = 3
FROM post p
WHERE p.id = 5
…

JPA merge – UPDATE statements
Query:[
"update post set title=? where id=?"
],
Params:[
(Vlad Mihalcea's High-Performance Java Persistence, Part no. 0, 1),
(Vlad Mihalcea's High-Performance Java Persistence, Part no. 2, 5)
]
Query:[
"update post_comment set post_id=?, review=? where id=?"
],
Params:[
(1, Excellent read!, 2),
(5, Excellent read!, 6)
]

Hibernate-specific update operation
Query:[
"update post set title=? where id=?"
],
Params:[
(Vlad Mihalcea's High-Performance Java Persistence, Part no. 2, 5)
]
Query:[
"update post_comment set post_id=?, review=? where id=?"
],
Params:[
(5, Excellent read!, 6)
]

Connections
CC BY 2.0 - https://www.flickr.com/photos/justinbaeder/5317820857/

Resource-local connection acquisition

Immediate connection acquisition
@PersistenceContext
private EntityManager entityManager;
@Transactional
public void importForecasts(String dataFilePath) {
Document forecastXmlDocument = readXmlDocument( dataFilePath );
List<Forecast> forecasts = parseForecasts(forecastXmlDocument);
for(Forecast forecast : forecasts) {
entityManager.persist( forecast );
}
}

Resource-local delay connection acquisition
<property
name="hibernate.connection.provider_disables_autocommit"
value="true"
/>
HikariConfig hikariConfig = new HikariConfig();
hikariConfig.setDataSourceClassName(dataSourceClassName);
hikariConfig.setDataSourceProperties(dataSourceProperties);
hikariConfig.setMinimumPoolSize(minPoolSize);
hikariConfig.setMaximumPoolSize(maxPoolSize);
hikariConfig.setAutoCommit(false);
DataSource datasource = new HikariDataSource(hikariConfig);

Resource-local connection acquisition optimization

Thank you
• Twitter: @vlad_mihalcea
• Blog: https://vladmihalcea.com
• Courses: https://vladmihalcea.com/courses
• Book: https://vladmihalcea.com/books/high-performance-java-persistence/

JPA and Hibernate Performance Tips

Recommended

More Related Content

What's hot (20)

Similar to JPA and Hibernate Performance Tips (20)

More from Vlad Mihalcea (7)

Recently uploaded (20)

JPA and Hibernate Performance Tips