Built-in Functions in LoraDB

Function names are case-insensitive; canonical camelCase is shown on each page. Unknown names and wrong arity are rejected at analysis time (Unknown function 'foo', WrongArity).

Most functions propagate null — any null argument makes the result null. The exceptions are the aggregates, coalesce, timestamp, pi, e, and rand.

Categories

Category	Examples	Reference
Aggregation	count, sum, collect, percentileCont	Aggregation
String	toLower, split, substring, replace	String
Math	abs, sqrt, sin, log, rand	Math
List	size, head, range, reduce	List
Temporal	date, datetime, duration.between	Temporal
Spatial	point, distance	Spatial
Vector	vector, vector.similarity.cosine, vector.similarity.euclidean, vector_distance, vector_norm, vector_dimension_count, toIntegerList, toFloatList	Vector
Path	length, nodes, relationships	Paths

Entity introspection

Function	Takes	Returns
id(x)	node | relationship	Int — internal id
labels(n)	node	List<String>
type(r)	relationship	String — rel type
keys(x)	node | rel | map	List<String>
properties(x)	node | rel | map	Map

MATCH (u:User)-[r:FOLLOWS]->(v:User)
RETURN id(u), labels(u), type(r), keys(u), properties(u)

Common uses

// Dump every property on a node as a map
MATCH (u:User {id: $id}) RETURN properties(u)

// Discover which labels a node carries
MATCH (n) WHERE id(n) = $raw_id RETURN labels(n)

// Inspect the type of a matched edge
MATCH (a)-[r]->(b) RETURN type(r), count(*) ORDER BY count(*) DESC

// Avoid duplicate pair rows
MATCH (a)-[:KNOWS]-(b) WHERE id(a) < id(b) RETURN a, b

See Graph model → Identity for why id() is opaque.

Type conversion and checking

Function	Behaviour
toString(x)	any → String; null → null
toInteger(x) / toInt(x)	Int/Float/String/Bool → Int or null
toFloat(x)	Int/Float/String → Float or null
toBoolean(x) / toBooleanOrNull(x)	Bool/String/Int → Bool or null
valueType(x)	name of the value's type, e.g. "INTEGER", "LIST<T>"
coalesce(a, b, …)	first non-null argument
timestamp()	current Unix time in milliseconds

RETURN toInteger('42'),                         -- 42
       toInteger('abc'),                        -- null
       toFloat(42),                             -- 42.0
       coalesce(null, null, 'fallback'),        -- 'fallback'
       valueType(1),                            -- 'INTEGER'
       valueType([1, 2, 3]),                    -- 'LIST<INTEGER>'
       valueType(date('2024-01-15'))            -- 'DATE'

coalesce recipes

// Default a missing property
MATCH (p:Person) RETURN p.name, coalesce(p.nickname, p.name) AS display

// Cascade through several optional fields
RETURN coalesce($phone, $email, 'unknown') AS contact

// Replace null in ordering
MATCH (p:Person)
RETURN p.name, coalesce(p.rank, 999999) AS rank_for_sort
ORDER BY rank_for_sort

For multi-branch logic with arbitrary predicates per branch (not just "first non-null"), use CASE.

valueType recipes

// Filter a heterogeneous list to numbers only
MATCH (n)
WHERE all(x IN n.values WHERE valueType(x) = 'INTEGER')
RETURN n

// Group by runtime type
UNWIND [1, 'two', 3.0, true, null] AS x
RETURN valueType(x) AS t, count(*) AS n
ORDER BY t

timestamp

Wall-clock milliseconds since the Unix epoch.

MERGE (c:Counter {name: 'events'})
  ON CREATE SET c.first_seen = timestamp()
  SET c.last_seen = timestamp()

See Data Types for every valueType return value and for how each type maps between LoraDB and host languages.

Null propagation — the common thread

Most functions return null when any argument is null. A small handful don't, so they're worth memorising:

• Aggregates (count, sum, …) skip null inputs (except count(*), which counts rows).
• coalesce(a, b, …) — returns the first non-null argument.
• timestamp(), pi(), e(), rand() — take no arguments.

Everywhere else, expect null in → null out. This is what makes IS NULL / IS NOT NULL essential over = null.

Quick lookup

Finding the right function for a task:

I want to…	Reach for
Pick the first non-null value	coalesce(a, b, …)
Branch on arbitrary conditions	CASE WHEN … THEN … END
Count rows matching a condition	count(CASE WHEN … THEN 1 END)
Concatenate a list into a string	reduce over split / collect
Current time (ms)	timestamp()
Current calendar day	date()
Name of a value's type	valueType(x)
Internal id of a node / rel	id(x)
Total order over temporal values	<, <=, >, >= — see Ordering
Cartesian or geodesic distance	distance(a, b)
Score a VECTOR against a query vector	vector.similarity.cosine(v, $q)
Signed distance under a metric	vector_distance(a, b, EUCLIDEAN)
Magnitude of a VECTOR	vector_norm(v, EUCLIDEAN)
Dimension of a VECTOR	vector_dimension_count(v) or size(v)
Convert VECTOR coordinates back to a LIST	toIntegerList(v) / toFloatList(v)

Not supported

• APOC-style utilities (apoc.*) — no compatibility layer.
• Procedures (CALL db.labels() etc.) — rejected at analysis time.
• User-defined functions — no registration surface.

Full list in Limitations.

See also

• Aggregation Functions — count, collect, percentiles.
• String, Math, List — everyday helpers.
• Temporal, Spatial — typed domains.
• Data Types Overview — value shapes.
• Queries → Parameters — binding typed values from the host.