Get graph schema

Description

The Get graph schema transform reads which node labels and relationship types a graph database has, and the properties of each. It writes one row per property of a label or relationship type, and one row without property for a label or type without properties.

Use it to document a graph, to check a graph against what a pipeline expects, or to see what a graph extracted with the Extract graph transform contains.

The transform doesn’t read input rows: it is the start of a pipeline, like Table input.

It works with every graph database connection and with Neo4j connections.

How the schema is read

Database Labels, types and properties Sample size

Database	Labels, types and properties	Sample size
Neo4j	The schema procedures `db.schema.nodeTypeProperties()` and `db.schema.relTypeProperties()`, which read the whole graph. The labels at the ends of the relationships come from `db.schema.visualization()`. Without these procedures, a sample.	Only for the sample
Memgraph	`SHOW SCHEMA INFO`, when Memgraph runs with `--schema-info-enabled`. Otherwise a sample.	Only for the sample
Amazon Neptune	A sample of the nodes and relationships	Yes
FalkorDB	The labels and relationship types from `db.labels()` and `db.relationshipTypes()`, the properties from a sample of every label and type	Per label and per type
Apache AGE	The labels and relationship types from the `ag_catalog.ag_label` table, the properties from a sample of every label and type	Per label and per type
Gremlin servers	The element maps of a sample of the vertices and edges	Yes

Neo4j

The schema procedures db.schema.nodeTypeProperties() and db.schema.relTypeProperties(), which read the whole graph. The labels at the ends of the relationships come from db.schema.visualization(). Without these procedures, a sample.

Only for the sample

Memgraph

SHOW SCHEMA INFO, when Memgraph runs with --schema-info-enabled. Otherwise a sample.

Only for the sample

Amazon Neptune

A sample of the nodes and relationships

Yes

FalkorDB

The labels and relationship types from db.labels() and db.relationshipTypes(), the properties from a sample of every label and type

Per label and per type

Apache AGE

The labels and relationship types from the ag_catalog.ag_label table, the properties from a sample of every label and type

Per label and per type

Gremlin servers

The element maps of a sample of the vertices and edges

Yes

A sample is the first nodes and the first relationships the database returns, or the first of every label and type where the database lists those. Properties which only occur outside the sample are missing, and a property is mandatory when every node or relationship in the sample has it.

The indexes come from the database too, where it lists them: Neo4j, Memgraph, FalkorDB and Apache AGE.

Options

Option Description

Option	Description
Transform name	The name of the transform, unique in the pipeline.
Graph database connection	The connection to read the schema of.
Sample size	The number of nodes and relationships to read, per label and relationship type where the database lists them. 1000 by default.
Element type field	The output field with `NODE` or `RELATIONSHIP`.
Name field	The output field with the node label or relationship type.
Property field	The output field with the property name, empty for a label or type without properties.
Property types field	The output field with the types of the property values, separated by commas. The type names are the same for every database: `String`, `Integer`, `Float`, `Boolean`, `Date`, `LocalDateTime`, `DateTime`, `LocalTime`, `Time`, `Duration`, `Point`, `Map`, `List` and `List<String>` and so on when all elements have the same type.
Mandatory field	The output field which is `true` if every node or relationship has the property, empty if unknown.
Indexed field	The output field which is `true` if an index covers the property, empty if the database doesn’t list its indexes.
Unique field	The output field which is `true` if a unique index or constraint covers the property, empty if unknown.
Start labels field	For relationship types: the output field with the labels of the nodes they start from, separated by commas.
End labels field	For relationship types: the output field with the labels of the nodes they end at, separated by commas.

Transform name

The name of the transform, unique in the pipeline.

Graph database connection

The connection to read the schema of.

Sample size

The number of nodes and relationships to read, per label and relationship type where the database lists them. 1000 by default.

Element type field

The output field with NODE or RELATIONSHIP.

Name field

The output field with the node label or relationship type.

Property field

The output field with the property name, empty for a label or type without properties.

Property types field

The output field with the types of the property values, separated by commas. The type names are the same for every database: String, Integer, Float, Boolean, Date, LocalDateTime, DateTime, LocalTime, Time, Duration, Point, Map, List and List<String> and so on when all elements have the same type.

Mandatory field

The output field which is true if every node or relationship has the property, empty if unknown.

Indexed field

The output field which is true if an index covers the property, empty if the database doesn’t list its indexes.

Unique field

The output field which is true if a unique index or constraint covers the property, empty if unknown.

Start labels field

For relationship types: the output field with the labels of the nodes they start from, separated by commas.

End labels field

For relationship types: the output field with the labels of the nodes they end at, separated by commas.

Leave the name of an output field empty to leave the field out.

Notes

A node with several labels counts for each of them. A property is mandatory for a label when every node with that label has it, whatever its other labels.