Spring Boot Starter CSV

Provides support for importing data by CSV files. This Spring Boot starter registers a CsvService and /csv endpoint, capable of handling CSV files.

It's also possible to import CSV files using a file directory.

Handler

Each type of CSV file demands a handler implementation. This handler describes how the file should be parsed and handled. Register a component that implements the CsvHandler interface:

@Slf4j
@Component
public class PersonCsvHandler implements CsvHandler<PersonCsvRow> {

  public static final String TYPE = "persons";

  @Override
  public String getType() {
    return TYPE;
  }

  @Override
  public Class<PersonCsvRow> getBeanClass() {
    return PersonCsvRow.class;
  }

  @Override
  public CsvResult handle(CsvClient<PersonCsvRow> client) {
    return new CsvTemplate().read(client::readBean, person -> save(person));
  }
  
}

Handlers are detected during application startup and provided to end users. When handlers should be protected, please add your own security logic.

Documentation

CSV files are documented by implementing the describe(CsvDocument) method in its handler:

@Slf4j
@Component
public class PersonCsvHandler implements CsvHandler<PersonCsvRow> {

  @Override
  public void describe(CsvDocument document) {
    document.setDescription("The persons known in this system.");
    document.addColumn(new CsvColumn("first_name").description("Person first name").example("Piet").required());
  }
  
}

CSV documentation is exposed by HTTP endpoint, allowing it be visualised while uploading.

To ensure the documentation remains valid we recommend writing tests. By calling the validate method a CSV upload is simulated using the examples from the document.

public class PersonCsvHandlerTest {

    @Test
    public void validate_shouldSucceed() {
        CsvResult result = csvService.validate(PersonCsvHandler.TYPE);
        Assertions.assertTrue(result.isSuccess());
        Assertions.assertEquals(1, result.getSuccess());
    }

}

HTTP endpoints

HTTP GET /csv

Retrieve all registered CSV types and default separator/quote characters. This endpoint is used to prefill the CSV upload form.

HTTP GET /csv/document?type=X

Retrieve the documentation of a specific CSV type. In the documentation we define the various columns, including an example CSV file.

HTTP POST /csv

Processes the CSV file based on the following required parameters:

• file: CSV file (multipart)
• type: CSV type name, as specified in handler
• separator: Separator character
• quote: Quote character

Content will be processed per line and results are returned as response body:

{
 "success": 10,
 "errors": [
  {
   "rowNumber": 2,
   "message": "Could not map column 'age' at index 4: For input string: 'not a number'"
  }
 ]
}

File share

Files can also be uploaded via the file system. Enabling file mode requires some configuration:

csv:
  file:
    cron: '*/5 * * * *' (cron when to scan)
    directory: /opt/app/csv (base directory)
    run_on_startup: true (scan on startup, default false)

In file mode the application will automatically scan the file system for new *.csv files in the upload directory of every registered handler and attempt to process them.

Files are moved to the work directory and processed. After processing it will either move to the success or fail directory. Error messages are moved in the logs directory.

The directory structure looks as follows:

 /opt/app/csv (base dir)
  /persons (csv type)
   /uploads (inbox)
   /work
   /success
   /fail
   /logs
  /orders (other type)

Make sure the user running the application has write access to the base directory. Sub directories will be created on demand.