Over a million developers have joined DZone.
Platinum Partner

URL shortener service in 42 lines of code in... Java (?!)

· IoT Zone

The IoT Zone is brought to you in partnership with ThingWorx. Explore the ThingWorx Developer Zone and learn how to rapidly develop IoT applications.

Apparently writing a URL shortener service is the new "Hello, world!" in the IoT/microservice/era world. It all started with A URL shortener service in 45 lines of Scala - neat piece of Scala, flavoured with Spray and Redis for storage. This was quickly followed with A url shortener service in 35 lines of Clojure and even URL Shortener in 43 lines of Haskell. So my inner anti-hipster asked: how long would it be in Java? But not plain Java, for goodness' sake. Spring Boot with Spring Data Redis are a good starting point. All we need is a simple controller handling GET and POST:

import com.google.common.hash.Hashing;
import org.apache.commons.validator.routines.UrlValidator;
import org.springframework.beans.factory.annotation.Autowired;
import org.springframework.boot.SpringApplication;
import org.springframework.data.redis.core.StringRedisTemplate;
import org.springframework.http.*;
import org.springframework.web.bind.annotation.*;

import javax.servlet.http.*;
import java.nio.charset.StandardCharsets;

public class UrlShortener {
    public static void main(String[] args) {
        SpringApplication.run(UrlShortener.class, args);

    @Autowired private StringRedisTemplate redis;

    @RequestMapping(value = "/{id}", method = RequestMethod.GET)
    public void redirect(@PathVariable String id, HttpServletResponse resp) throws Exception {
        final String url = redis.opsForValue().get(id);
        if (url != null)

    @RequestMapping(method = RequestMethod.POST)
    public ResponseEntity<String> save(HttpServletRequest req) {
        final String queryParams = (req.getQueryString() != null) ? "?" + req.getQueryString() : "";
        final String url = (req.getRequestURI() + queryParams).substring(1);
        final UrlValidator urlValidator = new UrlValidator(new String[]{"http", "https"});
        if (urlValidator.isValid(url)) {
            final String id = Hashing.murmur3_32().hashString(url, StandardCharsets.UTF_8).toString();
            redis.opsForValue().set(id, url);
            return new ResponseEntity<>("http://mydomain.com/" + id, HttpStatus.OK);
        } else
            return new ResponseEntity<>(HttpStatus.BAD_REQUEST);
The code is nicely self-descriptive and is functionally equivalent to a version in Scala. I didn't try to it squeeze too much to keep line count as short as possible, code above is quite typical with few details:

  • I don't normally use wildcard imports
  • I don't use fully qualified class names (I wanted to save one import line, I admit)
  • I surround if/else blocks with braces
  • I almost never use field injection, ugliest brother in inversion of control family. Instead I would go for constructor to allow testing with mocked Redis:

    private final StringRedisTemplate redis;
    public UrlShortener(StringRedisTemplate redis) {
        this.redis = redis;
The thing I struggled the most was... obtaining the original, full URL. Basically I needed everything after .com or port. No bloody way (neither servlets, nor Spring MVC), hence the awkward getQueryString() fiddling. You can use the service as follows - creating shorter URL:

$ curl -vX POST localhost:8080/https://www.google.pl/search?q=tomasz+nurkiewicz

> POST /https://www.google.pl/search?q=tomasz+nurkiewicz HTTP/1.1
> User-Agent: curl/7.30.0
> Host: localhost:8080
> Accept: */*
< HTTP/1.1 200 OK
< Server: Apache-Coyote/1.1
< Content-Type: text/plain;charset=ISO-8859-1
< Content-Length: 28
< Date: Sat, 23 Aug 2014 20:47:40 GMT
Redirecting through shorter URL:

$ curl -v localhost:8080/50784f51

> GET /50784f51 HTTP/1.1
> User-Agent: curl/7.30.0
> Host: localhost:8080
> Accept: */*
< HTTP/1.1 302 Found
< Server: Apache-Coyote/1.1
< Location: https://www.google.pl/search?q=tomasz+nurkiewicz
< Content-Length: 0
< Date: Sat, 23 Aug 2014 20:48:00 GMT
For completeness, here is a build file in Gradle (maven would work as well), skipped in all previous solutions:

buildscript {
    repositories {
        maven { url "http://repo.spring.io/libs-snapshot" }
    dependencies {
        classpath 'org.springframework.boot:spring-boot-gradle-plugin:1.1.5.RELEASE'

apply plugin: 'java'
apply plugin: 'spring-boot'

sourceCompatibility = '1.8'

repositories {
    maven { url 'http://repository.codehaus.org' }
    maven { url 'http://repo.spring.io/milestone' }

dependencies {
    compile "org.springframework.boot:spring-boot-starter-web:1.1.5.RELEASE"
    compile "org.springframework.boot:spring-boot-starter-redis:1.1.5.RELEASE"
    compile 'com.google.guava:guava:17.0'
    compile 'org.apache.commons:commons-lang3:3.3.2'
    compile 'commons-validator:commons-validator:1.4.0'
    compile 'org.apache.tomcat.embed:tomcat-embed-el:8.0.9'
    compile "org.aspectj:aspectjrt:1.8.1"

    runtime "cglib:cglib-nodep:3.1"

tasks.withType(GroovyCompile) {
    groovyOptions.optimizationOptions.indy = true

task wrapper(type: Wrapper) {
    gradleVersion = '2.0'
Actually also 42 lines... That's the whole application, no XML, no descriptors, not setup.

I don't treat this exercise as just a dummy code golf for shortest, most obfuscated working code. URL shortener web service with Redis back-end is an interesting showcase of syntax and capabilities of a given language and ecosystem. Much more entertaining then a bunch of algorithmic problems, e.g. found in Rosetta code. Also it's a good bare minimum template for writing a REST service.

One important feature of original Scala implementation, that was somehow silently forgotten in all implementations, including this one, is that it's non-blocking. Both HTTP and Redis access is event-driven (reactive, all right, I said it), thus I suppose it can handle tens of thousands of clients simultaneously. This can't be achieved with blocking controllers backed by Tomcat. But still you have to admit such a service written in Java (not even Java 8!) is surprisingly concise, easy to follow and straightforward - none of the other solutions are that readable (this is of course subjective).

Waiting for others!

The IoT Zone is brought to you in partnership with ThingWorx. ThingWorx is the most complete application development platform available that is purpose-built for the IoT.


Published at DZone with permission of Tomasz Nurkiewicz , DZone MVB .

Opinions expressed by DZone contributors are their own.

{{ parent.title || parent.header.title}}

{{ parent.tldr }}

{{ parent.urlSource.name }}