DEV Community

Cover image for CSV Generation From Large JSON Response in Spring
Ratul sharker
Ratul sharker

Posted on

CSV Generation From Large JSON Response in Spring

#java #tutorial #performance #webdev

Background

With the continuation of Streaming Large JSON Response in Spring and Consume Large JSON Response in Spring now time to put a real world scenario to generate a report based on the received response data. We will be generate a csv (comma separated value) report based on the received response data.

Goal

Consume Large JSON Response in Spring here we received list of Employee pojo which looks like following

Employee.java

public class Employee {

    private Long empNo;

    private Date birthDate;

    private String firstName;

    private String lastName;

    private Gender gender;

    private Date hireDate;
}
Enter fullscreen mode Exit fullscreen mode

From here we will be generating a csv file. The goal here to generate the whole csv file in the file system. Then serve the csv file over the network from the file system directly without putting pressure on the heap memory.

Firstly a get request will show a web page, containing the prompt for download the report. Clicking on the prompting button will initiate generating the csv file and end up downloading that file within the browser.

Implementation

First task is easy, designing a landing page with the "Download Report" prompt button.

inside the template folder

employee-report


... CSS details are skipped for focus

<body>
    <a href="/download-employee-report" download>Download Report</a>
</body>
Enter fullscreen mode Exit fullscreen mode

EmployeeController.java

@GetMapping("/employee-report")
public String getReportDownloadPage() {
    return "employee-report";
}
Enter fullscreen mode Exit fullscreen mode

Now running the consumer application, and hitting http://localhost:8081/employee-report will show up following

Download Report Prompt

Now time to implement route /download-employee-report.

Now we need the jackson-dataformat-csv dependency, using maven

pom.xml

<dependencies>
    <dependency>
        <groupId>com.fasterxml.jackson.dataformat</groupId>
        <artifactId>jackson-dataformat-csv</artifactId>
        <version>2.13.0</version>
    </dependency>
</dependencies>
Enter fullscreen mode Exit fullscreen mode

Before doing that we need CSVMapper bean

WebConfig.java

@Configuration
public class WebConfig {

    ... Previous Code

    @Bean
    @Primary
    public ObjectMapper restTemplateObjectMapper(RestTemplate restTemplate) ... Previous Code

    @Bean(name = "csv-mapper")
    public CsvMapper csvMapper() {
        CsvMapper csvMapper = new CsvMapper();        
        return csvMapper;
    }

}
Enter fullscreen mode Exit fullscreen mode

CSVMapper is subclass of ObjectMapper, so break the confusion while bean injection of ObjectMapper, we declared @Primary annotation on top of ObjectMapper bean creation. To inject the CSVMapper bean we need to add @Qualifier("csv-mapper") before bean injection.

Now Injecting the CSVMapper

EmployeeReportService.java

@Service
public class EmployeeReportService {

    private final RestTemplate restTemplate;
    private final ObjectMapper objectMapper;
    private final CsvMapper csvMapper;

    public EmployeeReportService(RestTemplate restTemplate, ObjectMapper objectMapper,
            @Qualifier("csv-mapper") CsvMapper csvMapper) {
        this.restTemplate = restTemplate;
        this.objectMapper = objectMapper;
        this.csvMapper = csvMapper;
    }

    ... Previous Code
}
Enter fullscreen mode Exit fullscreen mode

From Consume Large JSON Response in Spring we already have

EmployeeReportService.java

public Long fetchAllEmployee(Consumer<Employee> employeeConsumer) {

    String url = "http://localhost:8080/employees?stream=true";

    return restTemplate.execute(url, HttpMethod.GET, null, (response) -> {

        Long employeeCount = 0L;

        JsonParser jsonParser = objectMapper.getFactory().createParser(response.getBody());

        if(jsonParser.nextToken() == JsonToken.START_OBJECT) {
            if(jsonParser.nextFieldName() == "employees") {
                if(jsonParser.nextToken() == JsonToken.START_ARRAY) {
                    while(jsonParser.nextToken() != JsonToken.END_ARRAY) {
                        Employee employee = jsonParser.readValueAs(Employee.class);
                        employeeConsumer.accept(employee);
                        employeeCount++;
                    }
                }
            }
        }

        jsonParser.close();

        return employeeCount;
    });

}
Enter fullscreen mode Exit fullscreen mode

It allow us to getting Employee pojo one by one from the response. Now using the above method

EmployeeReportService.java

@Service
public class EmployeeReportService {

    ... Previous Code

    public File prepareEmployeeReportFile() throws IOException {

        CsvSchema schema = csvMapper.schemaFor(Employee.class);

        File tempFile = null;
        FileOutputStream fos = null;

        try {

            tempFile = File.createTempFile("employee", "csv");
            fos = new FileOutputStream(tempFile);
            CsvGenerator csvGenerator = csvMapper.getFactory().createGenerator(fos);
            csvGenerator.setSchema(schema);

            fetchAllEmployee((employee) -> {
                try {
                    csvGenerator.writeObject(employee);
                } catch(IOException ex) {
                    throw new RuntimeException(ex);
                }
            });


        } catch(IOException ex) {
            if(fos != null) {
                fos.close();
            }
            if(tempFile != null) {
                tempFile.delete();
            }
        } finally {
            if(fos != null) {
                fos.close();
            }
        }
        return tempFile;
    }
}
Enter fullscreen mode Exit fullscreen mode

From the top of the view, this method return the temporary csv File generated from the response.

Now explaining step by step.

  • First declare the schema for the csv file.
  • Create a temporary file into the file system's temp directory and getting the FileOutputStream.
  • From the CSVMapper bean and FileOutputStream from previous step, a CSVGenerator is prepared to write csv into the FileOutputStream. Schema generated in the first step, is also set on CSVGenerator.
  • Using the fetchAllEmployee writing Employee object one by one into the temporary file.
  • Whole task is wrapped inside try...catch...finally to ensure that any unwanted incident closes the FileOutputStream and deletes the temporary file.
  • Finally return the temporary file.

Now the /download-employee-report route implementation

EmployeeController.java

@GetMapping("/download-employee-report")
public void downloadEmployeeReport(HttpServletResponse response) throws IOException {

    response.setContentType("application/octet-stream");
    response.setHeader("Content-disposition", "attachment; filename=employees.csv");

    File csvTempFile = employeeReportService.prepareEmployeeReportFile();
    FileInputStream fis = new FileInputStream(csvTempFile);

    IOUtils.copy(fis, response.getOutputStream());

    fis.close();

    //TODO: Need to be more careful about deleting this file.
    csvTempFile.delete();
}
Enter fullscreen mode Exit fullscreen mode

Here, setting the proper headers for file content type and attachment file name. Copying the whole file into the HttpServletResponse's output stream. Thus ensured the whole file is not loaded into memory.

Finally closing the stream and remove the temporary file.

Here the implementation should be more careful, because any sort of IOException can occur during the IOUtils.copy(...).

Securing the implementation will look as follows

FileInputStream fis = null;
try {

    fis = new FileInputStream(csvTempFile);
    IOUtils.copy(fis, response.getOutputStream());

} catch(IOException ex) {
    // Handle the exception
} finally {
    if(fis != null) {
        fis.close();
    }

    if(csvTempFile != null) {
        csvTempFile.delete();
    }    
}
Enter fullscreen mode Exit fullscreen mode

Now running both the server and consumer application with -Xmx32m jvm argument, visit the http://localhost:8081/employee-report route and click the Download Report prompt button. It will initiate a employees.csv file download.

Content of the employees.csv look like following

-536824800000,10001,Georgi,M,504640800000,Facello
-189669600000,10002,Bezalel,F,473191200000,Simmel
-347522400000,10003,Parto,M,504640800000,Bamford
-505375200000,10004,Chirstian,M,504640800000,Koblick
-473925600000,10005,Kyoichi,M,599594400000,Maliniak
-536824800000,10006,Anneke,F,599594400000,Preusig
-410421600000,10007,Tzvetan,F,599594400000,Zielinski
-378972000000,10008,Saniya,M,756842400000,Kalloufi
... ~300K Entries
Enter fullscreen mode Exit fullscreen mode

So we successfully downloaded the csv file, but without header line nothing is understood.

To appear the header

EmployeeReportService.java

public File prepareEmployeeReportFile() throws IOException {

    CsvSchema schema = csvMapper.schemaFor(Employee.class).withHeader(); // Instruction for header to appear
    ... Previous Code     
}
Enter fullscreen mode Exit fullscreen mode

Now the output becomes

birthDate,empNo,firstName,gender,hireDate,lastName
-536824800000,10001,Georgi,M,504640800000,Facello
-189669600000,10002,Bezalel,F,473191200000,Simmel
-347522400000,10003,Parto,M,504640800000,Bamford
-505375200000,10004,Chirstian,M,504640800000,Koblick
-473925600000,10005,Kyoichi,M,599594400000,Maliniak
-536824800000,10006,Anneke,F,599594400000,Preusig
-410421600000,10007,Tzvetan,F,599594400000,Zielinski
-378972000000,10008,Saniya,M,756842400000,Kalloufi
... ~300K Entries
Enter fullscreen mode Exit fullscreen mode

Now the headers appear. But still few problems

  • Ordering of the columns are not as expected.
  • Dates are showing in in-human format.

By default the serial of the column is done in alphabetical order. To serial according to out necessity we need following on top of Employee.java

Employee.java

@Getter
@Setter
@NoArgsConstructor
@JsonPropertyOrder({"empNo", "firstName", "lastName", "gender", "birthDate", "hireDate"})
public class Employee {
    ... Previous Code
}
Enter fullscreen mode Exit fullscreen mode

Doing so the downloaded employees.csv looks like following

empNo,firstName,lastName,gender,birthDate,hireDate
10001,Georgi,Facello,M,-536824800000,504640800000
10002,Bezalel,Simmel,F,-189669600000,473191200000
10003,Parto,Bamford,M,-347522400000,504640800000
10004,Chirstian,Koblick,M,-505375200000,504640800000
10005,Kyoichi,Maliniak,M,-473925600000,599594400000
10006,Anneke,Preusig,F,-536824800000,599594400000
10007,Tzvetan,Zielinski,F,-410421600000,599594400000
10008,Saniya,Kalloufi,M,-378972000000,756842400000
... ~300K Entries
Enter fullscreen mode Exit fullscreen mode

So our column is according to our need.

Now time for the date format. To customise the date format update the CSVMapper bean creation

WebConfig.java

@Bean(name = "csv-mapper")
public CsvMapper csvMapper() {
    CsvMapper csvMapper = new CsvMapper();
    csvMapper.setDateFormat(new SimpleDateFormat("dd MMMM YYYY")); // Setting the date format.
    return csvMapper;
}
Enter fullscreen mode Exit fullscreen mode

Now the downloaded csv looks like following

empNo,firstName,lastName,gender,birthDate,hireDate
10001,Georgi,Facello,M,"28 December 1953","29 December 1986"
10002,Bezalel,Simmel,F,"29 December 1964","30 December 1985"
10003,Parto,Bamford,M,"28 December 1959","29 December 1986"
10004,Chirstian,Koblick,M,"27 December 1954","29 December 1986"
10005,Kyoichi,Maliniak,M,"26 December 1955","01 January 1989"
10006,Anneke,Preusig,F,"28 December 1953","01 January 1989"
10007,Tzvetan,Zielinski,F,"30 December 1957","01 January 1989"
... ~300K Entries
Enter fullscreen mode Exit fullscreen mode

So our expected date format is in action.

Still the header names are not human friendly. i.e instead of empNo "Employee No" is more preferred. To add this customisation, we need to add @JsonGetter in the getter method. We are using lombok, so adding the @JsonGetter will look like following

Employee.java

@Setter
@NoArgsConstructor
@JsonPropertyOrder({"empNo", "firstName", "lastName", "gender", "birthDate", "hireDate"})
public class Employee {

    @Getter(onMethod = @__(@JsonGetter("Employee No")))
    private Long empNo;

    @Getter(onMethod = @__(@JsonGetter("Birth Date")))
    private Date birthDate;

    @Getter(onMethod = @__(@JsonGetter("First Name")))
    private String firstName;

    @Getter(onMethod = @__(@JsonGetter("Last Name")))
    private String lastName;

    @Getter(onMethod = @__(@JsonGetter("Gender")))
    private Gender gender;

    @Getter(onMethod = @__(@JsonGetter("Hire Date")))
    private Date hireDate;
}
Enter fullscreen mode Exit fullscreen mode

Now trying with the csv will give us

"Employee No","First Name","Last Name",Gender,"Birth Date","Hire Date"
,,,,,
,,,,,
,,,,,
,,,,,
,,,,,
,,,,,
,,,,,
... ~300K Entries
Enter fullscreen mode Exit fullscreen mode

Header line is fine, but what happened to our data ?

Remember, the same pojo is used during deserialisation from the API. So we need to be more careful during deserialisation. We also need to update the @Setter with @JsonSetter. Doing so the final pojo will be

Employee.java

@NoArgsConstructor
@JsonPropertyOrder({"empNo", "firstName", "lastName", "gender", "birthDate", "hireDate"})
public class Employee {

    @Getter(onMethod = @__(@JsonGetter("Employee No")))
    @Setter(onMethod = @__(@JsonSetter("empNo")))
    private Long empNo;

    @Getter(onMethod = @__(@JsonGetter("Birth Date")))
    @Setter(onMethod = @__(@JsonSetter("birthDate")))
    private Date birthDate;

    @Getter(onMethod = @__(@JsonGetter("First Name")))
    @Setter(onMethod = @__(@JsonSetter("firstName")))
    private String firstName;

    @Getter(onMethod = @__(@JsonGetter("Last Name")))
    @Setter(onMethod = @__(@JsonSetter("lastName")))
    private String lastName;

    @Getter(onMethod = @__(@JsonGetter("Gender")))
    @Setter(onMethod = @__(@JsonSetter("gender")))
    private Gender gender;

    @Getter(onMethod = @__(@JsonGetter("Hire Date")))
    @Setter(onMethod = @__(@JsonSetter("hireDate")))
    private Date hireDate;
}
Enter fullscreen mode Exit fullscreen mode

Downloading the csv file will be look like

"Birth Date","Employee No","First Name",Gender,"Hire Date","Last Name"
"28 December 1953",10001,Georgi,M,"29 December 1986",Facello
"29 December 1964",10002,Bezalel,F,"30 December 1985",Simmel
"28 December 1959",10003,Parto,M,"29 December 1986",Bamford
"27 December 1954",10004,Chirstian,M,"29 December 1986",Koblick
"26 December 1955",10005,Kyoichi,M,"01 January 1989",Maliniak
"28 December 1953",10006,Anneke,F,"01 January 1989",Preusig
... ~300K Entries
Enter fullscreen mode Exit fullscreen mode

Now see the header line is scrambled again. Fixing the values of @JsonPropertyOrder will look like

@NoArgsConstructor
@JsonPropertyOrder({"Employee No", "First Name", "Last Name", "Gender", "Birth Date", "Hire Date"})
public class Employee {
    ... Previous Code
}
Enter fullscreen mode Exit fullscreen mode

Doing so the generated csv will be as follows

"Employee No","First Name","Last Name",Gender,"Birth Date","Hire Date"
10001,Georgi,Facello,M,"28 December 1953","29 December 1986"
10002,Bezalel,Simmel,F,"29 December 1964","30 December 1985"
10003,Parto,Bamford,M,"28 December 1959","29 December 1986"
10004,Chirstian,Koblick,M,"27 December 1954","29 December 1986"
10005,Kyoichi,Maliniak,M,"26 December 1955","01 January 1989"
10006,Anneke,Preusig,F,"28 December 1953","01 January 1989"
... ~300K Entries
Enter fullscreen mode Exit fullscreen mode

Perfect !!!

Still few things are not quiet right, downloading of the csv took long time, but there is no indication that we have clicked the "Download Report" and csv generation is in progress. This issue will be covered in a separate post.

All the code above can be found into my github repository.

Top comments (0)

Read next

madetocode profile image

Beyond the Calculator: Exploring Reverse Polish Notation with JavaScript

Moyomade -

aurelievache profile image

Understanding Go: part 10 – Struct

Aurélie Vache -

railsdesigner profile image

ViewComponent over Turbo Stream Broadcasts

Rails Designer -

ashsajal profile image

JavaScript Event Loop Explained in 1 minute.

Ashfiquzzaman Sajal -