Object Relational Mapping (ORM) in computer science is a programming technique for converting data between incompatible type systems in object-oriented programming languages.

In object-oriented programming, data-management tasks act on object-oriented (OO) objects that are almost always non-scalar values.

For example, consider an address book entry that represents a single person along with zero or more phone numbers and zero or more addresses. This could be modeled in an object-oriented implementation by a "Person object" with attributes/fields to hold each data item that the entry comprises: the person's name, a list of phone numbers, and a list of addresses. The list of phone numbers would itself contain "PhoneNumber objects" and so on. The address-book entry is treated as a single object by the programming language (it can be referenced by a single variable containing a pointer to the object, for instance). Various methods can be associated with the object, such as a method to return the preferred phone number, the home address, and so on.

However, many popular database products such as SQL database management systems (DBMS) can only store and manipulate scalar values such as integers and strings organized within tables. The programmer must either convert the object values into groups of simpler values for storage in the database (and convert them back upon retrieval), or only use simple scalar values within the program. Object-relational mapping implements the first approach.

The heart of the problem involves translating the logical representation of the objects into an atomized form that is capable of being stored in the database, while preserving the properties of the objects and their relationships so that they can be reloaded as objects when needed. If this storage and retrieval functionality is implemented, the objects are said to be persistent.

Persistence

ORM is concerned with helping your application to achieve persistence.

Persistence means that we would like our application’s data to outlive the applications process. In Java terms, we would like the state of (some of) our objects to live beyond the scope of the JVM so that the same state is available later.

Relational Databases

Specifically, ORM is concerned with data persistence as it applies to relational databases (RDBMS). RDBMS is a very popular persistence mechanism.

The Object-Relational Impedance Mismatch

'Object-Relational Impedance Mismatch' (sometimes called the 'paradigm mismatch') is a way of saying that object models and relational models do not work very well together. RDBMSs represent data in a tabular format, whereas object-oriented languages, such as Java, represent it as objects. Loading and storing objects using a tabular relational database exposes us to 5 mismatch problems…​

Granularity

Sometimes you will have an object model which has more classes than the number of corresponding tables in the database, we says the object model is more granular than the relational model.

Subtypes (Inheritance)

Inheritance is a natural paradigm in object-oriented programming languages. However, RDBMSs do not define anything similar on the whole.

Identity

A RDBMS defines exactly one notion of 'sameness': the primary key. Java, however, defines both object identity a==b and object equality a.equals(b).

Associations

Associations are represented as unidirectional references in Object Oriented languages whereas RDBMSs use the notion of foreign keys. If you need bidirectional relationships in Java, you must define the association twice.

Likewise, you cannot determine the multiplicity of a relationship by looking at the object domain model.

Data navigation

The way you access data in Java is fundamentally different than the way you do it in a relational database. In Java, you navigate from one association to another walking the object network.

This is not an efficient way of retrieving data from a relational database. You typically want to minimize the number of SQL queries and thus load several entities via JOINs and select the targeted entities before you start walking the object network.

Object-relational mapping