Yorm is a basic ORM-alike framework designed to work with Java Records, without class generation, neither annotations. In the world of microservices, there is a tendency to have very contained logic within every service, and hence reduced databases, that in many cases are simply no more than several tables with not that many fields. Java Records usually are a perfect fit for basic CRUD operations, and here is where Yorm shines.
Yorm needs at least Java 17, although it works of newer versions as well.
Yorm might be for you in case:
- You are working with microservices with Java and like Java Records
- Your relational databases are pretty simple and basic
- Your tables contain auto increment ids
- You don't need complex INNER JOIN queries, just basic CRUD
Although the Java industry offers very well maintained ORM solutions like Hibernate or Jooq, they tend not to work that well with Java Records. Yorm on the other side is specifically designed to leverage this Java capability.
Due to the immutable nature of Java Records, Yorm cannot be understood as a persistent ORM, not even as an ORM, as there aren't really any relationships.
- No need to generate classes
- No need to add annotations
- No need to write SQL for basic operations
- Seamless flow with API REST and CRUD operations
Yorm doesn't need to generate classes or to annotate them, but it works on conventions. It will assume that your table has a Primary Key called id, probably with an autoincrement. Also it will assume that foreign keys will follow the naming patter of table_id. The convention will assume as well that fields in the table and fields in the record will have the same name, or a very similar one.
When a Java Record is operated with Yorm, a reflection inspection will came in, and all the methods of the Record will be matched with their counterparts from the database. This matching will be kept in memory as a map, to avoid using reflection again. However, if there is a change in the database structure, the microservice will probably need to be restarted to refresh this mapping.
Yorm has been designed to need very few dependencies:
- HikariCP - Hikari, to deal with the database
- MySql - Database supported and tested
- PostgreSql - Database supported and tested, courtesy of PabloGrisafi
- Junit 5 - For the unit tests
- TestContainers - Also for the unit tests
- Slf4j - Logging is usually useful
And that's it, the Yorm lies heavily on Java Records and Reflections.
Yorm has been tested with Snowflake, and it works as long as the table has no primary keys, and the Jvm is run with the option --add-opens java.base/java.nio=ALL-UNNAMED
Imagine you have a database with a table called Person, defined like:
CREATE TABLE person
(
id INT(10) AUTO_INCREMENT PRIMARY KEY NOT NULL,
name VARCHAR(20) NOT NULL,
email VARCHAR(55) NOT NULL,
company_id int(10) NOT NULL
);
Hence, you a Java Record that can match that table:
public record Person(int id, String name, String email, int companyId) {}
Please note how companyId follows the traditional camelcase in Java, and company_id the underscore which is very popular in the database world. It can be like that, Yorm will take care and match both fields.
Saving an object will be something as easy as:
Person person = new Person(0, "John", "john.doe@um.com", 1);
int idPerson = yorm.save(person);
This will translate to SQL:
INSERT INTO person (name, email, company_id) VALUES ("John", "john.doe@um.com", 1)
Yorm will detect that the id has a 0, consider it an INSERT, and insert it in the database, in table Person. It will return the id of the object just inserted.
We might need to insert the object with its id, there is no problem:
Person person = new Person(2, "Mark", "mark.doe@um.com", 1);
int idPerson = yorm.insert(person);
This operation will be automatically translated into
INSERT INTO person (id, name, email, company_id) VALUES (2, "Mark", "mark.doe@um.com", 1)
The update operation follows the same pattern. Please bear in mind that Records are immutable, so we have to create a new one, and the id will be used to detect that it's an update operation:
Person person = new Person(2, "Jacob", "jacob.doe@um.com", 1);
yorm.save(person);
Whose equivalent SQL would be:
UPDATE person SET name = "Jacob", email="jacob.doe@um.com" WHERE id=2 AND company_id=1
Why is using company_id here? Well, it looks like a candidate to be a foreign key to a company table:
CREATE TABLE company
(
id INT(10) AUTO_INCREMENT PRIMARY KEY NOT NULL,
name VARCHAR(20) NOT NULL,
country_code VARCHAR(2) NOT NULL,
creation_date DATE NOT NULL,
is_active TINYINT NOT NULL
);
ALTER TABLE `person`
ADD CONSTRAINT `person_id` FOREIGN KEY (`company_id`) REFERENCES `company` (`id`) ON DELETE CASCADE;
Yorm will detect that there is one Primary Key and a Multiple Key, and use them for the update. Nevertheless, Yorm also allows the update operation:
Person person = new Person(2, "Jacob", "jacob.doe@um.com", 1);
yorm.update(person);
Insertion can even be massive:
Person person1 = new Person(2, "Hermione", "hermione.granger@hogwarts.com", 1);
Person person2 = new Person(3, "Harry", "harry.potter@hogwarts.com", 1);
Person person3 = new Person(4, "Sauron", "sauron@mordor.com", 2);
List<Person> list = List.of(person1, person2, person3);
yorm.insert(list);
We've inserted and updated elements in the table. How can we retrieve them into Records? The first and easiest way would be retrieving all the elements. It's just one line:
List<Person> personList = yorm.find(Person.class);
Which is the same as the SQL:
SELECT id, name, email, company_id FROM person
And automatically wrapping the result into a List. Of course, we could just get one element, if we know the id:
Person person = yorm.find(Person.class, 1);
Which translates to SQL of:
SELECT id, name, email, company_id FROM person WHERE id=1
Or even retrieve elements with a foreign key:
Company company = new Company(1, null, null, null, false);
List<Person> personList = yorm.find(Person.class, company);
Pay attention how the Record company, with id 1, is used to retrieve a list of objects Person whose company_id is 1. Yorm will detect that Person has a foreign key with Company and use that to build the query:
SELECT id, name, email, company_id FROM person WHERE company_id=1
Some kinds of filtering is also allowed:
Person personFilter1 = new Person(0, "harry", "john", 0);
Person personFilter2 = new Person(0, null, null, 2);
List<Person> list = List.of(personFilter1, personFilter2);
List<Person> personList = yorm.find(list);
Let's review this one a bit. Yorm deals with Records. Here we have defined two Person records, the first one with some information on name and email, and the second one with a company_id. Yorm will ignore all the nulls and 0s on fields that are ids, but use the rest of the information to build a SELECT query and retrieve a List of Person Records:
SELECT id, name, email, company_id FROM person WHERE name like '%harry%' OR email like '%john%' OR company_id=2
Based on Benjiql idea, Yorm also has some sort of fluent API capabilities to find records. The previous example could also be written like:
List<Person> personList = yorm.from(Person.class).where(Person::name).like("harry")
.or(Person::email).like("john")
.or(Person::companyId).equalTo(2)
.find();
Yorm can chain operands to build SQL alike sentences just with code:
List<Person> thirdList = yorm.from(Person.class).where(Person::name).equalTo("John")
.and(Person::lastLogin).greaterThan(LocalDateTime.of(2019, 01, 01, 0, 0, 0))
.find();
Which translated to SQL would be:
SELECT id, name, email, last_login, company_id FROM person WHERE name = 'John' AND last_login >= '2019-01-01 00:00'
As a final note, Yorm works just by creating an instance of Yorm with a javax.sql.DataSource:
DataSource ds = DbConnector.getDatasource(parameters);
Yorm yorm = new Yorm(ds);
Yorm is very young but nevertheless quite useful if you just want to perform basic CRUD operations, specially the ones that involve REST endpoints in the world of microservices. It's very to useful, and very transparent, since you just need a DataSource to put it to work.
You can easily add Yorm to your project with
<dependency>
<groupId>org.yorm</groupId>
<artifactId>yorm</artifactId>
<version>0.0.8</version>
</dependency>
- Docker
Yorm is a small library published in Maven Central. It can be very easily compiled like
mvn clean install
Apache 2.0