PGX 20.1.1
Documentation
Documentation
The distributed execution mode supports Graph Pattern Matching through a subset of PGQL 1.1 and PGQL 1.2.
In summary, the PGX distributed execution mode supports:
| Supported Feature | PGQL Version | Example |
|---|---|---|
| Expression support (with projections, GROUP BY, ORDER BY) | 1.1 | SELECT 42 + AVG(m.salary) MATCH (n)->(m) GROUP BY m ORDER BY AVG(m.salary) % 10 |
| Pattern matching -- directed or any-directed edge patterns -- with filters (and labels) | 1.1 | SELECT ... MATCH (a:person)-[e:likes]->(b:post) WHERE a.age > 18 AND a.group = b.group (directed); or SELECT ... MATCH (a:person)-[e:follows]-(b:person) (any-directed) |
Aggregations without GROUP BY |
1.1 | SELECT COUNT(*), MAX(a.age), AVG(a.age) MATCH (a:person)-[e:likes]->(b:post) WHERE a.age > 18 |
GROUP BY |
1.1 | SELECT a.age, AVG(a.income) ... GROUP BY a.age |
ORDER BY (with LIMIT and OFFSET) |
1.1 | SELECT a.age, ... ORDER BY a.age |
Explicit casting (CAST) |
1.1 | SELECT CAST(n.age AS FLOAT), CAST(n.salary AS INT), CAST('True' AS BOOLEAN), CAST(TRUE AS STRING) MATCH (n) |
Distinct projections (SELECT DISTINCT) |
1.1 | SELECT DISTINCT n.age, n.salary MATCH (n) |
Pattern matching with IN filter expression |
1.2 | SELECT ... MATCH (a:person)-[e:likes]->(b:post) WHERE a.name IN ('Emily', 'Paul', 'Juan') |
Field extraction from temporal values (EXTRACT) |
1.2 | SELECT EXTRACT(YEAR FROM DATE '2017-02-13'), EXTRACT(SECOND FROM TIME '12:05:03.201') MATCH (n) |
Numeric functions (ABS, CEIL/CEILING, FLOOR, ROUND) |
1.2 | SELECT ABS(-5), FLOOR(5.7), CEIL(5.2), ROUND(5.5) MATCH (n) |
CASE statements |
1.2 | SELECT CASE n.name WHEN 'John' THEN n.age ELSE n.income END MATCH (n) |
Filtering of groups (HAVING) |
1.2 | SELECT a, SUM(a.income) MATCH (a) GROUP BY a HAVING SUM(a.income) > 1000 |
The PGX distributed execution mode does not support any other PGQL 1.2 feature (also described here). Additionally, it does not support the following PGQL 1.1 features:
| Unsupported Feature | Example |
|---|---|
| Disconnected patterns (Cartesian product) | SELECT COUNT(a) MATCH (a), (b) |
| Subqueries | SELECT COUNT(a) MATCH (a) WHERE a.age < (SELECT AVG(b.age) MATCH (b)) |
| Regular path queries | SELECT COUNT(a) MATCH (a) -/:likes+/-> (b) |
IS NULL and IS NOT NULL functions |
SELECT a MATCH (a) WHERE a.address IS NULL |
all_different(v1, v2, v3, ..) and has_prop(v, prop) built-in functions |
SELECT a, b, c MATCH (a)->(b)-(c) WHERE all_diferent(a, b, c) |
| Build-time constant expressions in combination with aggregations in projections | SELECT 1 * 3 + 4, AVG(a) MATCH (a) GROUP BY a |
The system will report error messages if unsupported queries are issued.