JarCastingPgBulkInsert
PgBulkInsert is a Java library for Bulk Inserts to PostgreSQL using the Binary COPY Protocol.
It provides a wrapper around the PostgreSQL COPY command:
The COPY command is a PostgreSQL specific feature, which allows efficient bulk import or export of data to and from a table. This is a much faster way of getting data in and out of a table than using INSERT and SELECT.
This project wouldn't be possible without the great Npgsql library, which has a beautiful implementation of the Postgres protocol.
Setup
PgBulkInsert is available in the Central Maven Repository.
You can add the following dependencies to your pom.xml to include PgBulkInsert in your project.
<dependency>
<groupId>de.bytefish</groupId>
<artifactId>pgbulkinsert</artifactId>
<version>8.0.0</version>
</dependency>
Supported PostgreSQL Types
- Numeric Types
- smallint
- integer
- bigint
- real
- double precision
- numeric
- Date/Time Types
- timestamp
- timestamptz
- date
- time
- Character Types
- text
- JSON Types
- jsonb
- Boolean Type
- boolean
- Binary Data Types
- bytea
- Network Address Types
- inet (IPv4, IPv6)
- macaddr
- UUID Type
- uuid
- Array Type
- One-Dimensional Arrays
- Range Type
- int4range
- int8range
- numrange
- tsrange
- tstzrange
- daterange
- hstore
- hstore
- Geometric Types
- point
- line
- lseg
- box
- path
- polygon
- circle
Usage
You can use the PgBulkInsert API in various ways. The first one is to use the SimpleRowWriter when you don't have an explicit Java POJO, that matches a Table. The second way is to use an AbstractMapping<TEntityType> to define a mapping between a Java POJO and a PostgreSQL table.
Please also read the FAQ, which may answer some of your questions.
Using the SimpleRowWriter
Using the SimpleRowWriter doesn't require you to define a separate mapping. It requires you to define the PostgreSQL table structure using a SimpleRowWriter.Table, that has a schema name (optional), table name and column names:
// Schema of the Table:
String schemaName = "sample";
// Name of the Table:
String tableName = "row_writer_test";
// Define the Columns to be inserted:
String[] columnNames = new String[] {
"value_int",
"value_text"
};
// Create the Table Definition:
SimpleRowWriter.Table table = new SimpleRowWriter.Table(schemaName, tableName, columnNames);
Once created you create the SimpleRowWriter by using the Table and a PGConnection.
Now to write a row to PostgreSQL you call the startRow method. It expects you to pass a Consumer<SimpleRow> into it, which defines what data to write to the row. The call to startRow is synchronized, so it is safe to be called from multiple threads.
// Create the Writer:
try(SimpleRowWriter writer = new SimpleRowWriter(table, pgConnection)) {
// ... write your data rows:
for(int rowIdx = 0; rowIdx < 10000; rowIdx++) {
// ... using startRow and work with the row, see how the order doesn't matter:
writer.startRow((row) -> {
row.setText("value_text", "Hi");
row.setInteger("value_int", 1);
});
}
}
So the complete example looks like this:
public class SimpleRowWriterTest extends TransactionalTestBase {
// ...
@Test
public void rowBasedWriterTest() throws SQLException {
// Get the underlying PGConnection:
PGConnection pgConnection = PostgreSqlUtils.getPGConnection(connection);
// Schema of the Table:
String schemaName = "sample";
// Name of the Table:
String tableName = "row_writer_test";
// Define the Columns to be inserted:
String[] columnNames = new String[] {
"value_int",
"value_text"
};
// Create the Table Definition:
SimpleRowWriter.Table table = new SimpleRowWriter.Table(schemaName, tableName, columnNames);
// Create the Writer:
try(SimpleRowWriter writer = new SimpleRowWriter(table, pgConnection)) {
// ... write your data rows:
for(int rowIdx = 0; rowIdx < 10000; rowIdx++) {
// ... using startRow and work with the row, see how the order doesn't matter:
writer.startRow((row) -> {
row.setText("value_text", "Hi");
row.setInteger("value_int", 1);
});
}
}
// Now assert, that we have written 10000 entities:
Assert.assertEquals(10000, getRowCount());
}
}
If you need to customize the Null Character Handling, then you can use the setNullCharacterHandler(Function<String, String> nullCharacterHandler) function.
Using the AbstractMapping
The AbstractMapping is the second possible way to map a POJO for usage in PgBulkInsert. Imagine we want to bulk insert a large amount of people into a PostgreSQL database. Each Person has a first name, a last name and a birthdate.
Database Table
The table in the PostgreSQL database might look like this:
CREATE TABLE sample.unit_test
(
first_name text,
last_name text,
birth_date date
);
Domain Model
The domain model in the application might look like this:
private class Person {
private String firstName;
private String lastName;
private LocalDate birthDate;
public Person() {}
public String getFirstName() {
return firstName;
}
public void setFirstName(String firstName) {
this.firstName = firstName;
}
public String getLastName() {
return lastName;
}
public void setLastName(String lastName) {
this.lastName = lastName;
}
public LocalDate getBirthDate() {
return birthDate;
}
public void setBirthDate(LocalDate birthDate) {
this.birthDate = birthDate;
}
}
Bulk Inserter
Then you have to implement the AbstractMapping<Person>, which defines the mapping between the table and the domain model.
public class PersonMapping extends AbstractMapping<Person>
{
public PersonMapping() {
super("sample", "unit_test");
mapText("first_name", Person::getFirstName);
mapText("last_name", Person::getLastName);
mapDate("birth_date", Person::getBirthDate);
}
}
This mapping is used to create the PgBulkInsert<Person>:
PgBulkInsert<Person> bulkInsert = new PgBulkInsert<Person>(new PersonMapping());
Using the Bulk Inserter
And finally we can write a Unit Test to insert 100000 people into the database. You can find the entire Unit Test on GitHub as IntegrationTest.java.
@Test
public void bulkInsertPersonDataTest() throws SQLException {
// Create a large list of People:
List<Person> personList = getPersonList(100000);
// Create the BulkInserter:
PgBulkInsert<Person> bulkInsert = new PgBulkInsert<Person>(new PersonMapping(schema));
// Now save all entities of a given stream:
bulkInsert.saveAll(PostgreSqlUtils.getPGConnection(connection), personList.stream());
// And assert all have been written to the database:
Assert.assertEquals(100000, getRowCount());
}
private List<Person> getPersonList(int num) {
List<Person> personList = new ArrayList<>();
for (int pos = 0; pos < num; pos++) {
Person p = new Person();
p.setFirstName("Philipp");
p.setLastName("Wagner");
p.setBirthDate(LocalDate.of(1986, 5, 12));
personList.add(p);
}
return personList;
}
FAQ
How can I write Primitive Types (boolean, float, double)?
By default methods like mapBoolean map the boxed type Boolean, Integer, Long. This might be problematic if you need to squeeze out the last seconds when doing bulk inserts, see Issue:
So for every data type that also has a primitive type, you can add a "Primitive" suffix to the method name like:
- ```mapBooleanPrimitive``
This will use the primitive type and prevent boxing and unboxing of values.
How can I write a java.sql.Timestamp?
You probably have Java classes with a java.sql.Timestamp in your application. Now if you use the AbstractMapping or a SimpleRowWriter it expects a LocalDateTime. Here is how to map a java.sql.Timestamp.
Imagine you have an EMail class with a property emailCreateTime, that is using a java.sql.Timestamp to represent the time. The column name in Postgres is email_create_time and you are using a timestamp data type.
To map the java.sql.Timestamp you would write the mapTimeStamp method like this:
mapTimeStamp("email_create_time", x -> x.getEmailCreateTime() != null ? x.getEmailCreateTime().toLocalDateTime() : null);
And here is the complete example:
public class EMail {
private Timestamp emailCreateTime;
public Timestamp getEmailCreateTime() {
return emailCreateTime;
}
}
public static class EMailMapping extends AbstractMapping<EMail>
{
public EMailMapping(String schema) {
super(schema, "unit_test");
mapTimeStamp("email_create_time", x -> x.getEmailCreateTime() != null ? x.getEmailCreateTime().toLocalDateTime() : null);
}
}
Handling Null Characters or... 'invalid byte sequence for encoding "UTF8": 0x00'
If you see the error message invalid byte sequence for encoding "UTF8": 0x00 your data contains Null Characters. Although 0x00 is totally valid UTF-8... PostgreSQL does not support writing it, because it uses C-style string termination internally.
PgBulkInsert allows you to enable a Null Value handling, that removes all 0x00 occurences and replaces them with an empty string:
// Create the Table Definition:
SimpleRowWriter.Table table = new SimpleRowWriter.Table(schema, tableName, columnNames);
// Create the Writer:
SimpleRowWriter writer = new SimpleRowWriter(table);
// Enable the Null Character Handler:
writer.enableNullCharacterHandler();
Running the Tests
Running the Tests requires a PostgreSQL database.
You have to configure the test database connection in the module pgbulkinsert-core and file db.properties:
db.url=jdbc:postgresql://127.0.0.1:5432/sampledb
db.user=philipp
db.password=test_pwd
db.schema=public
The tests are transactional, that means any test data will be rolled back once a test finishes. But it probably makes sense to set up a separate db.schema for your tests, if you want to avoid polluting the public schema or have different permissions.
License
PgBulkInsert is released with under terms of the MIT License: