This guide walks you through the process of creating a basic batch-driven solution.
You will build a service that imports data from a CSV spreadsheet, transforms it with custom code, and stores the final results in a database.
Typically, your customer or a business analyst supplies a spreadsheet. For this simple
example, you can find some made-up data in src/main/resources/sample-data.csv
:
link:initial/src/main/resources/sample-data.csv[role=include]
This spreadsheet contains a first name and a last name on each row, separated by a comma. This is a fairly common pattern that Spring can handle without customization.
Next, you need to write an SQL script to create a table to store the data. You can find
such a script in src/main/resources/schema-all.sql
:
link:initial/src/main/resources/schema-all.sql[role=include]
Note
|
Spring Boot runs schema-@@platform@@.sql automatically during startup. -all is the default for all platforms.
|
You can use this pre-initialized project and click Generate to download a ZIP file. This project is configured to fit the examples in this tutorial.
To manually initialize the project:
-
Navigate to https://start.spring.io. This service pulls in all the dependencies you need for an application and does most of the setup for you.
-
Choose either Gradle or Maven and the language you want to use. This guide assumes that you chose Java.
-
Click Dependencies and select Spring Batch and HyperSQL Database.
-
Click Generate.
-
Download the resulting ZIP file, which is an archive of a web application that is configured with your choices.
Note
|
If your IDE has the Spring Initializr integration, you can complete this process from your IDE. |
Note
|
You can also fork the project from Github and open it in your IDE or other editor. |
Now that you can see the format of data inputs and outputs, you can write code to
represent a row of data, as the following example (from
src/main/java/com/example/batchprocessing/Person.java
) shows:
link:complete/src/main/java/com/example/batchprocessing/Person.java[role=include]
You can instantiate the Person
class either with first and last name through a
constructor or by setting the properties.
A common paradigm in batch processing is to ingest data, transform it, and then pipe it
out somewhere else. Here, you need to write a simple transformer that converts the names
to uppercase. The following listing (from
src/main/java/com/example/batchprocessing/PersonItemProcessor.java
) shows how to do
so:
link:complete/src/main/java/com/example/batchprocessing/PersonItemProcessor.java[role=include]
PersonItemProcessor
implements Spring Batch’s ItemProcessor
interface. This makes it
easy to wire the code into a batch job that you will define later in this guide. According
to the interface, you receive an incoming Person
object, after which you transform it to
an upper-cased Person
.
Note
|
The input and output types need not be the same. In fact, after one source of data is read, sometimes the application’s data flow needs a different data type. |
Now you need to put together the actual batch job. Spring Batch provides many utility classes that reduce the need to write custom code. Instead, you can focus on the business logic.
To configure your job, you must first create a Spring @Configuration
class like the following example in
src/main/java/com/exampe/batchprocessing/BatchConfiguration.java
:
link:complete/src/main/java/com/example/batchprocessing/BatchConfiguration.java[role=include]
...
}
For starters, the @EnableBatchProcessing
annotation adds many critical beans that
support jobs and save you a lot of leg work. This example uses a memory-based database
(provided by @EnableBatchProcessing
), meaning that, when it is done, the data is gone. It also autowires a couple factories needed further below.
Now add the following beans to your BatchConfiguration
class to define a reader, a processor, and a writer:
link:complete/src/main/java/com/example/batchprocessing/BatchConfiguration.java[role=include]
The first chunk of code defines the input, processor, and output.
-
reader()
creates anItemReader
. It looks for a file calledsample-data.csv
and parses each line item with enough information to turn it into aPerson
. -
processor()
creates an instance of thePersonItemProcessor
that you defined earlier, meant to convert the data to upper case. -
writer(DataSource)
creates anItemWriter
. This one is aimed at a JDBC destination and automatically gets a copy of the dataSource created by@EnableBatchProcessing
. It includes the SQL statement needed to insert a singlePerson
, driven by Java bean properties.
The last chunk (from src/main/java/com/example/batchprocessing/BatchConfiguration.java
)
shows the actual job configuration:
link:complete/src/main/java/com/example/batchprocessing/BatchConfiguration.java[role=include]
The first method defines the job, and the second one defines a single step. Jobs are built from steps, where each step can involve a reader, a processor, and a writer.
In this job definition, you need an incrementer, because jobs use a database to maintain execution state. You then list each step, (though this job has only one step). The job ends, and the Java API produces a perfectly configured job.
In the step definition, you define how much data to write at a time. In this case, it writes up to ten records at a time. Next, you configure the reader, processor, and writer by using the beans injected earlier.
Note
|
chunk() is prefixed <Person,Person> because it is a generic method. This
represents the input and output types of each “chunk” of processing and lines up with
ItemReader<Person> and ItemWriter<Person> .
|
The last bit of batch configuration is a way to get notified when the job completes. The
following example (from
src/main/java/com/example/batchprocessing/JobCompletionNotificationListener.java
) shows
such a class:
link:/complete/src/main/java/com/example/batchprocessing/JobCompletionNotificationListener.java[role=include]
The JobCompletionNotificationListener
listens for when a job is BatchStatus.COMPLETED
and then uses JdbcTemplate
to inspect the results.
Although batch processing can be embedded in web apps and WAR files, the simpler approach demonstrated below creates a standalone application. You package everything in a single, executable JAR file, driven by a good old Java main()
method.
The Spring Initializr created an application class for you. For this simple example, it
works without further modification. The following listing (from
src/main/java/com/example/batchprocessing/BatchProcessingApplication.java
) shows the
application class:
link:complete/src/main/java/com/example/batchprocessing/BatchProcessingApplication.java[role=include]
Note that SpringApplication.exit()
and System.exit()
ensure that the JVM exits upon job completion.
See the Application Exit section in Spring Boot Reference documentation for more details.
For demonstration purposes, there is code to create a JdbcTemplate
, query the database,
and print out the names of people the batch job inserts.
The job prints out a line for each person that gets transformed. After the job runs, you can also see the output from querying the database. It should resemble the following output:
Converting (firstName: Jill, lastName: Doe) into (firstName: JILL, lastName: DOE)
Converting (firstName: Joe, lastName: Doe) into (firstName: JOE, lastName: DOE)
Converting (firstName: Justin, lastName: Doe) into (firstName: JUSTIN, lastName: DOE)
Converting (firstName: Jane, lastName: Doe) into (firstName: JANE, lastName: DOE)
Converting (firstName: John, lastName: Doe) into (firstName: JOHN, lastName: DOE)
Found <firstName: JILL, lastName: DOE> in the database.
Found <firstName: JOE, lastName: DOE> in the database.
Found <firstName: JUSTIN, lastName: DOE> in the database.
Found <firstName: JANE, lastName: DOE> in the database.
Found <firstName: JOHN, lastName: DOE> in the database.
Congratulations! You built a batch job that ingested data from a spreadsheet, processed it, and wrote it to a database.