Skip to content

Entity Relationship Diagrams ​

An entity–relationship model (or ER model) describes interrelated things of interest in a specific domain of knowledge. A basic ER model is composed of entity types (which classify the things of interest) and specifies relationships that can exist between entities (instances of those entity types). Wikipedia.

Note that practitioners of ER modelling almost always refer to entity types simply as entities. For example the CUSTOMER entity type would be referred to simply as the CUSTOMER entity. This is so common it would be inadvisable to do anything else, but technically an entity is an abstract instance of an entity type, and this is what an ER diagram shows - abstract instances, and the relationships between them. This is why entities are always named using singular nouns.

Mermaid can render ER diagrams

Entity names are often capitalised, although there is no accepted standard on this, and it is not required in Mermaid.

Relationships between entities are represented by lines with end markers representing cardinality. Mermaid uses the most popular crow's foot notation. The crow's foot intuitively conveys the possibility of many instances of the entity that it connects to.

ER diagrams can be used for various purposes, ranging from abstract logical models devoid of any implementation details, through to physical models of relational database tables. It can be useful to include attribute definitions on ER diagrams to aid comprehension of the purpose and meaning of entities. These do not necessarily need to be exhaustive; often a small subset of attributes is enough. Mermaid allows them to be defined in terms of their type and name.

When including attributes on ER diagrams, you must decide whether to include foreign keys as attributes. This probably depends on how closely you are trying to represent relational table structures. If your diagram is a logical model which is not meant to imply a relational implementation, then it is better to leave these out because the associative relationships already convey the way that entities are associated. For example, a JSON data structure can implement a one-to-many relationship without the need for foreign key properties, using arrays. Similarly an object-oriented programming language may use pointers or references to collections. Even for models that are intended for relational implementation, you might decide that inclusion of foreign key attributes duplicates information already portrayed by the relationships, and does not add meaning to entities. Ultimately, it's your choice.

Syntax ​

Entities and Relationships ​

Mermaid syntax for ER diagrams is compatible with PlantUML, with an extension to label the relationship. Each statement consists of the following parts:

    <first-entity> [<relationship> <second-entity> : <relationship-label>]


  • first-entity is the name of an entity. Names must begin with an alphabetic character or an underscore (from v10.5.0+), and may also contain digits and hyphens.
  • relationship describes the way that both entities inter-relate. See below.
  • second-entity is the name of the other entity.
  • relationship-label describes the relationship from the perspective of the first entity.

For example:

    PROPERTY ||--|{ ROOM : contains

This statement can be read as a property contains one or more rooms, and a room is part of one and only one property. You can see that the label here is from the first entity's perspective: a property contains a room, but a room does not contain a property. When considered from the perspective of the second entity, the equivalent label is usually very easy to infer. (Some ER diagrams label relationships from both perspectives, but this is not supported here, and is usually superfluous).

Only the first-entity part of a statement is mandatory. This makes it possible to show an entity with no relationships, which can be useful during iterative construction of diagrams. If any other parts of a statement are specified, then all parts are mandatory.

Relationship Syntax ​

The relationship part of each statement can be broken down into three sub-components:

  • the cardinality of the first entity with respect to the second,
  • whether the relationship confers identity on a 'child' entity
  • the cardinality of the second entity with respect to the first

Cardinality is a property that describes how many elements of another entity can be related to the entity in question. In the above example a PROPERTY can have one or more ROOM instances associated to it, whereas a ROOM can only be associated with one PROPERTY. In each cardinality marker there are two characters. The outermost character represents a maximum value, and the innermost character represents a minimum value. The table below summarises possible cardinalities.

Value (left)Value (right)Meaning
|oo|Zero or one
||||Exactly one
}oo{Zero or more (no upper limit)
}||{One or more (no upper limit)


Value (left)Value (right)Alias for
one or zeroone or zeroZero or one
zero or onezero or oneZero or one
one or moreone or moreOne or more
one or manyone or manyOne or more
many(1)many(1)One or more
1+1+One or more
zero or morezero or moreZero or more
zero or manyzero or manyZero or more
many(0)many(1)Zero or more
0+0+Zero or more
only oneonly oneExactly one
11Exactly one

Identification ​

Relationships may be classified as either identifying or non-identifying and these are rendered with either solid or dashed lines respectively. This is relevant when one of the entities in question can not have independent existence without the other. For example a firm that insures people to drive cars might need to store data on NAMED-DRIVERs. In modelling this we might start out by observing that a CAR can be driven by many PERSON instances, and a PERSON can drive many CARs - both entities can exist without the other, so this is a non-identifying relationship that we might specify in Mermaid as: PERSON }|..|{ CAR : "driver". Note the two dots in the middle of the relationship that will result in a dashed line being drawn between the two entities. But when this many-to-many relationship is resolved into two one-to-many relationships, we observe that a NAMED-DRIVER cannot exist without both a PERSON and a CAR - the relationships become identifying and would be specified using hyphens, which translate to a solid line:


ValueAlias for
optionally tonon-identifying

Attributes ​

Attributes can be defined for entities by specifying the entity name followed by a block containing multiple type name pairs, where a block is delimited by an opening { and a closing }. The attributes are rendered inside the entity boxes. For example:

The type values must begin with an alphabetic character and may contain digits, hyphens, underscores, parentheses and square brackets. The name values follow a similar format to type, but may start with an asterisk as another option to indicate an attribute is a primary key. Other than that, there are no restrictions, and there is no implicit set of valid data types.

Entity Name Aliases (v10.5.0+) ​

An alias can be added to an entity using square brackets. If provided, the alias will be showed in the diagram instead of the entity name.

Attribute Keys and Comments ​

Attributes may also have a key or comment defined. Keys can be PK, FK or UK, for Primary Key, Foreign Key or Unique Key. To specify multiple key constraints on a single attribute, separate them with a comma (e.g., PK, FK).. A comment is defined by double quotes at the end of an attribute. Comments themselves cannot have double-quote characters in them.

Other Things ​

  • If you want the relationship label to be more than one word, you must use double quotes around the phrase
  • If you don't want a label at all on a relationship, you must use an empty double-quoted string

Styling ​

Config options ​

For simple color customization:

NameUsed as
fillBackground color of an entity or attribute
strokeBorder color of an entity or attribute, line color of a relationship

Classes used ​

The following CSS class selectors are available for richer styling:

.er.attributeBoxEvenThe box containing attributes on even-numbered rows
.er.attributeBoxOddThe box containing attributes on odd-numbered rows
.er.entityBoxThe box representing an entity
.er.entityLabelThe label for an entity
.er.relationshipLabelThe label for a relationship
.er.relationshipLabelBoxThe box surrounding a relationship label
.er.relationshipLineThe line representing a relationship between entities