Supported PGQL Features and Limitations

This section provides the complete list of supported and unsupported PGQL functionalities in PGQL queries that can be performed directly on the PGQL and SQL property graphs in the database and those that are run after loading the graphs into memory.

Feature	PGQL on RDBMS	PGQL on RDBMS	PGQL on the Graph Server (PGX)
	(PGQL Property Graph)	(SQL Property GraphFoot 1)
CREATE PROPERTY GRAPH	Supported	Supported	Supported Limitations: No composite keys for vertices Properties need to be column references; arbitrary property expressions are not supported unless the graph is first created in the database and then loaded into the graph server (PGX).
DROP PROPERTY GRAPH	Supported	Supported	Not Supported
Fixed-length pattern matching	Supported	Supported	Supported
Variable-length pattern matching goals	Supported: Reachability Path search prefixes: ANY ANY SHORTEST SHORTEST k ALL Path modes: WALK TRAIL SIMPLE ACYCLIC Limitations: Path search prefixes: ALL SHORTEST ANY CHEAPEST CHEAPEST k	Not Supported	Supported: Reachability Path search prefixes: ANY ANY SHORTEST SHORTEST k ALL SHORTEST ANY CHEAPEST CHEAPEST k ALL Path modes: WALK TRAIL SIMPLE ACYCLIC
Variable-length pattern matching quantifiers	Supported: * + ? { n } { n, } { n, m } { , m }	Not Supported	Supported: * + ? { n } { n, } { n, m } { , m } Limitations: ? is only supported for reachability In case of ANY CHEAPEST and TOP k CHEAPEST, only * is supported
Variable-length path unnesting	Supported: ONE ROW PER STEP Limitation: Quantifier * not supported Not supported: ONE ROW PER VERTEX	Not Supported	Supported: ONE ROW PER VERTEX ONE ROW PER STEP Limitation: * quantifier is not supported
GROUP BY	Supported	Supported	Supported
HAVING	Supported	Supported	Supported
Aggregations	Supported: COUNT MIN, MAX, AVG, SUM LISTAGG JSON_ARRAYAGG Limitations: ARRAY_AGG	Supported: COUNT MIN, MAX, AVG, SUM LISTAGG Not supported: ARRAY_AGG JSON_ARRAYAGG	Supported: COUNT MIN, MAX, AVG, SUM LISTAGG ARRAY_AGG Not Supported: JSON_ARRAYAGG
DISTINCT SELECT DISTINCT Aggregation with DISTINCT (such as, COUNT(DISTINCT e.prop))	Supported	Supported	Supported
SELECT v.*	Supported	Not Supported	Supported
ORDER BY (+ASC/DESC), LIMIT, OFFSET	Supported	Supported	Supported
Data Types	All available Oracle RDBMS data types supported	All available Oracle RDBMS data types supported	Supported: INTEGER (32-bit) LONG (64-bit) FLOAT (32-bit) DOUBLE (64-bit) STRING (no maximum length) BOOLEAN DATE TIME TIME WITH TIME ZONE TIMESTAMP TIMESTAMP WITH TIME ZONE
JSON	Supported: JSON storage: JSON strings (VARCHAR2) JSON objects JSON functions: Any JSON function call that follows the syntax, json_function_name(arg1, arg2,…). For example: json_value(department_data, '$.department') Limitations: Simple Dot Notation Any optional clause in a JSON function call (such as RETURNING, ERROR, and so on) is not supported. For example: json_value(department_data, '$.employees[1].hireDate' RETURNING DATE)	Supported: JSON storage: JSON strings (VARCHAR2) JSON objects JSON functions: Any JSON function call that follows the syntax, json_function_name(arg1, arg2,…). For example: json_value(department_data, '$.department') Limitations: Simple Dot Notation Any optional clause in a JSON function call (such as RETURNING, ERROR, and so on) is not supported. For example: json_value(department_data, '$.employees[1].hireDate' RETURNING DATE)	No built-in JSON support. However, JSON values can be stored as STRING and manipulated or queried through user-defined functions (UDFs) written in Java or JavaScript.
Operators	Supported: Relational: +, -, *, /, %, - (unary minus) Arithmetic: =, <>, <, >, <=, >= Logical: AND, OR, NOT String: || (concat)	Supported: Relational: +, -, *, /, %, - (unary minus) Arithmetic: =, <>, <, >, <=, >= Logical: AND, OR, NOT String: || (concat)	Supported: Relational: +, -, *, /, %, - (unary minus) Arithmetic: =, <>, <, >, <=, >= Logical: AND, OR, NOT String: || (concat)
Functions and predicates	Supported are all available functions in the Oracle RDBMS that take the form function_name(arg1, arg2, ...) with optional schema and package qualifiers. Supported PGQL functions/predicates: IS NULL, IS NOT NULL JAVA_REGEXP_LIKE (based on CONTAINS) LOWER, UPPER SUBSTRING ABS, CEIL/CEILING, FLOOR, ROUND EXTRACT ID LABEL, HAS_LABEL ALL_DIFFERENT CAST CASE IN and NOT IN Limitations: LABELS IN_DEGREE, OUT_DEGREE	Supported are all available functions in the Oracle RDBMS that take the form function_name(arg1, arg2, ...) with optional schema and package qualifiers. Supported PGQL functions/predicates: IS NULL, IS NOT NULL LOWER, UPPER SUBSTRING ABS, CEIL/CEILING, FLOOR, ROUND EXTRACT CAST CASE IN and NOT IN Unsupported PGQL functions/predicates are all vertex/edge functions	Supported: IS NULL, IS NOT NULL JAVA_REGEXP_LIKE (based on CONTAINS) LOWER, UPPER SUBSTRING ABS, CEIL/CEILING, FLOOR, ROUND EXTRACT ID, VERTEX_ID, EDGE_ID LABEL, LABELS, IS [NOT] LABELED ALL_DIFFERENT IN_DEGREE, OUT_DEGREE CAST CASE IN and NOT IN MATCHNUM ELEMENT_NUMBER IS [NOT] SOURCE [OF], IS [NOT] DESTINATION [OF] VERTEX_EQUAL, EDGE_EQUAL
User-defined functions	Supported: PL/SQL functions Functions created via the Oracle Database Multilingual Engine (MLE)	Supported: PL/SQL functions Functions created via the Oracle Database Multilingual Engine (MLE)	Supported: Java UDFs JavaScript UDFs
Subqueries: Scalar subqueries EXISTS and NOT EXISTS subqueries LATERAL subquery	Supported: EXISTS and NOT EXISTS subqueries Scalar subqueries LATERAL subquery	Supported subqueries: EXISTS NOT EXISTS Not supported: Scalar subqueries LATERAL subquery	Supported
GRAPH_TABLE operator	Supported Extension: BASE GRAPHS clause in CREATE PROPERTY GRAPH for creating graphs based on metadata of other graphs	Not supported	Supported
INSERT/UPDATE/DELETE	Supported	Not supported	Supported
INTERVAL literals and operations	Not supported	Not supported	Supported literals: SECOND MINUTE HOUR DAY MONTH YEAR Supported operations: Add INTERVAL to datetime (+) Subtract INTERVAL from datetime (-)

^Footnote 1 SQL Property Graphs are supported only in Oracle Database 23ai, currently available with Always Free Autonomous Database.