fix(pg): Inconsistent conditional pattern matching BED-6695

cypher/models/pgsql/test/translation_cases/nodes.sql

@@ -208,13 +208,13 @@ with s0 as (select (n0.id, n0.kind_ids, n0.properties)::nodecomposite as n0 from
 with s0 as (select (n0.id, n0.kind_ids, n0.properties)::nodecomposite as n0 from node n0) select s0.n0 as s from s0 where (not exists (select 1 from edge e0 where e0.start_id = (s0.n0).id or e0.end_id = (s0.n0).id));
 
 -- case: match (s) where not (s)-[]->()-[]->() return s
-with s0 as (select (n0.id, n0.kind_ids, n0.properties)::nodecomposite as n0 from node n0) select s0.n0 as s from s0 where (not (with s1 as (select s0.n0 as n0, (n1.id, n1.kind_ids, n1.properties)::nodecomposite as n1 from s0 join edge e0 on (s0.n0).id = e0.start_id join node n1 on n1.id = e0.end_id), s2 as (select s1.n0 as n0, s1.n1 as n1 from s1 join edge e1 on (s1.n1).id = e1.start_id join node n2 on n2.id = e1.end_id) select count(*) > 0 from s2));
+with s0 as (select (n0.id, n0.kind_ids, n0.properties)::nodecomposite as n0 from node n0) select s0.n0 as s from s0 where (not (with s1 as (select s0.n0 as n0, (n1.id, n1.kind_ids, n1.properties)::nodecomposite as n1 from edge e0 join node n1 on n1.id = e0.end_id where (s0.n0).id = e0.start_id), s2 as (select s1.n0 as n0, s1.n1 as n1 from s1 join edge e1 on (s1.n1).id = e1.start_id join node n2 on n2.id = e1.end_id) select count(*) > 0 from s2));
 
 -- case: match (s) where not (s)-[{prop: 'a'}]-({name: 'n3'}) return s
 with s0 as (select (n0.id, n0.kind_ids, n0.properties)::nodecomposite as n0 from node n0) select s0.n0 as s from s0 where (not (with s1 as (select s0.n0 as n0 from s0 join edge e0 on ((s0.n0).id = e0.end_id or (s0.n0).id = e0.start_id) join node n1 on ((n1.properties -> 'name'))::jsonb = to_jsonb(('n3')::text)::jsonb and (n1.id = e0.end_id or n1.id = e0.start_id) where ((s0.n0).id <> n1.id) and ((e0.properties -> 'prop'))::jsonb = to_jsonb(('a')::text)::jsonb) select count(*) > 0 from s1));
 
 -- case: match (s) where not (s)<-[{prop: 'a'}]-({name: 'n3'}) return s
-with s0 as (select (n0.id, n0.kind_ids, n0.properties)::nodecomposite as n0 from node n0) select s0.n0 as s from s0 where (not (with s1 as (select s0.n0 as n0 from s0 join edge e0 on (s0.n0).id = e0.end_id join node n1 on ((n1.properties -> 'name'))::jsonb = to_jsonb(('n3')::text)::jsonb and n1.id = e0.start_id where ((e0.properties -> 'prop'))::jsonb = to_jsonb(('a')::text)::jsonb) select count(*) > 0 from s1));
+with s0 as (select (n0.id, n0.kind_ids, n0.properties)::nodecomposite as n0 from node n0) select s0.n0 as s from s0 where (not (with s1 as (select s0.n0 as n0 from edge e0 join node n1 on ((n1.properties -> 'name'))::jsonb = to_jsonb(('n3')::text)::jsonb and n1.id = e0.start_id where ((e0.properties -> 'prop'))::jsonb = to_jsonb(('a')::text)::jsonb and (s0.n0).id = e0.end_id) select count(*) > 0 from s1));
 
 -- case: match (n:NodeKind1) where n.distinguishedname = toUpper('admin') return n
 with s0 as (select (n0.id, n0.kind_ids, n0.properties)::nodecomposite as n0 from node n0 where (((n0.properties -> 'distinguishedname'))::jsonb = to_jsonb((upper('admin')::text)::text)::jsonb) and n0.kind_ids operator (pg_catalog.@>) array [1]::int2[]) select s0.n0 as n from s0;
@@ -229,7 +229,7 @@ with s0 as (select (n0.id, n0.kind_ids, n0.properties)::nodecomposite as n0 from
 with s0 as (select (n0.id, n0.kind_ids, n0.properties)::nodecomposite as n0 from node n0 where (cypher_ends_with((n0.properties ->> 'distinguishedname'), (upper('admin')::text)::text)::bool) and n0.kind_ids operator (pg_catalog.@>) array [1]::int2[]) select s0.n0 as n from s0;
 
 -- case: match (s) where not (s)-[{prop: 'a'}]->({name: 'n3'}) return s
-with s0 as (select (n0.id, n0.kind_ids, n0.properties)::nodecomposite as n0 from node n0) select s0.n0 as s from s0 where (not (with s1 as (select s0.n0 as n0 from s0 join edge e0 on (s0.n0).id = e0.start_id join node n1 on ((n1.properties -> 'name'))::jsonb = to_jsonb(('n3')::text)::jsonb and n1.id = e0.end_id where ((e0.properties -> 'prop'))::jsonb = to_jsonb(('a')::text)::jsonb) select count(*) > 0 from s1));
+with s0 as (select (n0.id, n0.kind_ids, n0.properties)::nodecomposite as n0 from node n0) select s0.n0 as s from s0 where (not (with s1 as (select s0.n0 as n0 from edge e0 join node n1 on ((n1.properties -> 'name'))::jsonb = to_jsonb(('n3')::text)::jsonb and n1.id = e0.end_id where ((e0.properties -> 'prop'))::jsonb = to_jsonb(('a')::text)::jsonb and (s0.n0).id = e0.start_id) select count(*) > 0 from s1));
 
 -- case: match (s) where not (s)-[]-() return id(s)
 with s0 as (select (n0.id, n0.kind_ids, n0.properties)::nodecomposite as n0 from node n0) select (s0.n0).id from s0 where (not exists (select 1 from edge e0 where e0.start_id = (s0.n0).id or e0.end_id = (s0.n0).id));
@@ -338,3 +338,6 @@ with s0 as (select (n0.id, n0.kind_ids, n0.properties)::nodecomposite as n0 from
 
 -- case: match (n) where n.name = "alpha' || (SELECT inet_server_addr()::text::int) || '" return n
 with s0 as (select (n0.id, n0.kind_ids, n0.properties)::nodecomposite as n0 from node n0 where (((n0.properties -> 'name'))::jsonb = to_jsonb(('alpha'' || (SELECT inet_server_addr()::text::int) || ''')::text)::jsonb)) select s0.n0 as n from s0;
+
+-- case: match (g:NodeKind2) where not ((g)<-[:EdgeKind1]-(:NodeKind1)) return g
+with s0 as (select (n0.id, n0.kind_ids, n0.properties)::nodecomposite as n0 from node n0 where n0.kind_ids operator (pg_catalog.@>) array [2]::int2[]) select s0.n0 as g from s0 where (not ((with s1 as (select s0.n0 as n0 from edge e0 join node n1 on n1.kind_ids operator (pg_catalog.@>) array [1]::int2[] and n1.id = e0.start_id where e0.kind_id = any (array [3]::int2[]) and (s0.n0).id = e0.end_id) select count(*) > 0 from s1)));

cypher/models/pgsql/test/translation_cases/stepwise_traversal.sql

@@ -98,7 +98,7 @@ with s0 as (select (n0.id, n0.kind_ids, n0.properties)::nodecomposite as n0, (n1
 with s0 as (select (n0.id, n0.kind_ids, n0.properties)::nodecomposite as n0, (n1.id, n1.kind_ids, n1.properties)::nodecomposite as n1 from edge e0 join node n0 on n0.kind_ids operator (pg_catalog.@>) array [1]::int2[] and n0.id = e0.start_id join node n1 on n1.kind_ids operator (pg_catalog.@>) array [2]::int2[] and n1.id = e0.end_id where e0.kind_id = any (array [3, 4]::int2[])) select ((s0.n0).properties -> 'name'), ((s0.n1).properties -> 'name') from s0;
 
 -- case: match (s)-[r:EdgeKind1]->() where (s)-[r {prop: 'a'}]->() return s
-with s0 as (select (e0.id, e0.start_id, e0.end_id, e0.kind_id, e0.properties)::edgecomposite as e0, (n0.id, n0.kind_ids, n0.properties)::nodecomposite as n0 from edge e0 join node n0 on n0.id = e0.start_id join node n1 on n1.id = e0.end_id where ((e0.properties -> 'prop'))::jsonb = to_jsonb(('a')::text)::jsonb and e0.kind_id = any (array [3]::int2[])) select s0.n0 as s from s0 where ((with s1 as (select s0.e0 as e0, s0.n0 as n0 from s0 join edge e0 on (s0.n0).id = (s0.e0).start_id join node n2 on n2.id = (s0.e0).end_id) select count(*) > 0 from s1));
+with s0 as (select (e0.id, e0.start_id, e0.end_id, e0.kind_id, e0.properties)::edgecomposite as e0, (n0.id, n0.kind_ids, n0.properties)::nodecomposite as n0 from edge e0 join node n0 on n0.id = e0.start_id join node n1 on n1.id = e0.end_id where ((e0.properties -> 'prop'))::jsonb = to_jsonb(('a')::text)::jsonb and e0.kind_id = any (array [3]::int2[])) select s0.n0 as s from s0 where ((with s1 as (select s0.e0 as e0, s0.n0 as n0 from edge e0 join node n2 on n2.id = (s0.e0).end_id where (s0.n0).id = (s0.e0).start_id) select count(*) > 0 from s1));
 
 -- case: match (s)-[r:EdgeKind1]->(e) where not (s.system_tags contains 'admin_tier_0') and id(e) = 1 return id(s), labels(s), id(r), type(r)
 with s0 as (select (e0.id, e0.start_id, e0.end_id, e0.kind_id, e0.properties)::edgecomposite as e0, (n0.id, n0.kind_ids, n0.properties)::nodecomposite as n0, (n1.id, n1.kind_ids, n1.properties)::nodecomposite as n1 from edge e0 join node n1 on (n1.id = 1) and n1.id = e0.end_id join node n0 on (not (coalesce((n0.properties ->> 'system_tags'), '')::text like '%admin\_tier\_0%')) and n0.id = e0.start_id where e0.kind_id = any (array [3]::int2[])) select (s0.n0).id, (array(select _kind.name from generate_subscripts((s0.n0).kind_ids, 1) as _kind_idx, kind _kind where _kind.id = ((s0.n0).kind_ids)[_kind_idx] order by _kind_idx))::text[], (s0.e0).id, kind_name((s0.e0).kind_id)::text from s0;

cypher/models/pgsql/translate/predicate.go

@@ -83,6 +83,8 @@ func (s *Translator) translatePatternPredicate() error {
 	return nil
 }
 
+// buildPatternPredicates is used by translateMatch to resolve deferred pattern predicate
+// futures collected for the current MATCH/OPTIONAL MATCH query part's WHERE expressions
 func (s *Translator) buildPatternPredicates() error {
 	for _, predicateFuture := range s.query.CurrentPart().patternPredicates {
 		var (
@@ -142,7 +144,17 @@ func (s *Translator) buildPatternPredicates() error {
 					})
 				}
 			} else {
-				if traversalStepQuery, err := s.buildTraversalPatternRoot(traversalStep.Frame, traversalStep); err != nil {
+				var (
+					traversalStepQuery pgsql.Query
+					err                error
+				)
+				if traversalStep.Direction != graph.DirectionBoth && (traversalStep.LeftNodeBound || traversalStep.RightNodeBound) {
+					traversalStepQuery, err = s.buildTraversalPatternRootWithOuterCorrelation(traversalStep.Frame, traversalStep)
+				} else {
+					traversalStepQuery, err = s.buildTraversalPatternRoot(traversalStep.Frame, traversalStep)
+				}
+
+				if err != nil {
 					return err
 				} else {
 					subQuery.AddCTE(pgsql.CommonTableExpression{

cypher/models/pgsql/translate/traversal.go

@@ -51,7 +51,8 @@ func (s *Translator) buildDirectionlessTraversalPatternRoot(traversalStep *Trave
 				},
 				JoinOperator: pgsql.JoinOperator{
 					JoinType:   pgsql.JoinTypeInner,
-					Constraint: pgsql.OptionalAnd(rightJoinLocal, traversalStep.RightNodeJoinCondition)},
+					Constraint: pgsql.OptionalAnd(rightJoinLocal, traversalStep.RightNodeJoinCondition),
+				},
 			}},
 		})
 
@@ -138,6 +139,116 @@ func (s *Translator) buildDirectionlessTraversalPatternRoot(traversalStep *Trave
 	}, nil
 }
 
+// buildTraversalPatternRootWithOuterCorrelation constructs a traversal pattern root, preserving the correlation to
+// the outer query part's context
+func (s *Translator) buildTraversalPatternRootWithOuterCorrelation(partFrame *Frame, traversalStep *TraversalStep) (pgsql.Query, error) {
+	if traversalStep.Direction == graph.DirectionBoth {
+		return s.buildDirectionlessTraversalPatternRoot(traversalStep)
+	}
+
+	var (
+		// Partition right-node constraints: only locally-scoped terms go into JOIN ON.
+		// Constraints that reference comma-connected CTEs (e.g. s0.i0 from a prior WITH)
+		// must remain in WHERE — they are out of scope inside an explicit JOIN chain.
+		rightJoinLocal, rightJoinExternal = partitionConstraintByLocality(
+			traversalStep.RightNodeConstraints,
+			pgsql.AsIdentifierSet(traversalStep.RightNode.Identifier, traversalStep.Edge.Identifier),
+		)
+
+		nextSelect = pgsql.Select{
+			Projection: traversalStep.Projection,
+		}
+	)
+
+	if traversalStep.LeftNodeBound && traversalStep.RightNodeBound {
+		nextSelect.From = append(nextSelect.From, pgsql.FromClause{
+			Source: pgsql.TableReference{
+				Name:    pgsql.CompoundIdentifier{pgsql.TableEdge},
+				Binding: models.OptionalValue(traversalStep.Edge.Identifier),
+			},
+		})
+
+		// Both nodes of the traversal are fully bound by the outer query and the frame bindings
+		// will have been rewritten to reference the outer CTEs here, so we don't need any JOINs
+		// and can use those conditions inside of the inner WHERE to correlate the result set.
+		nextSelect.Where = pgsql.OptionalAnd(traversalStep.LeftNodeConstraints, nextSelect.Where)
+		nextSelect.Where = pgsql.OptionalAnd(traversalStep.RightNodeConstraints, nextSelect.Where)
+		nextSelect.Where = pgsql.OptionalAnd(traversalStep.LeftNodeJoinCondition, nextSelect.Where)
+		nextSelect.Where = pgsql.OptionalAnd(traversalStep.RightNodeJoinCondition, nextSelect.Where)
+		nextSelect.Where = pgsql.OptionalAnd(traversalStep.EdgeConstraints.Expression, nextSelect.Where)
+
+		return pgsql.Query{
+			Body: nextSelect,
+		}, nil
+	} else if traversalStep.LeftNodeBound {
+		nextSelect.From = append(nextSelect.From, pgsql.FromClause{
+			Source: pgsql.TableReference{
+				Name:    pgsql.CompoundIdentifier{pgsql.TableEdge},
+				Binding: models.OptionalValue(traversalStep.Edge.Identifier),
+			},
+			Joins: []pgsql.Join{{
+				Table: pgsql.TableReference{
+					Name:    pgsql.CompoundIdentifier{pgsql.TableNode},
+					Binding: models.OptionalValue(traversalStep.RightNode.Identifier),
+				},
+				JoinOperator: pgsql.JoinOperator{
+					JoinType:   pgsql.JoinTypeInner,
+					Constraint: pgsql.OptionalAnd(rightJoinLocal, traversalStep.RightNodeJoinCondition),
+				},
+			}},
+		})
+
+		nextSelect.Where = pgsql.OptionalAnd(traversalStep.LeftNodeConstraints, nextSelect.Where)
+		nextSelect.Where = pgsql.OptionalAnd(traversalStep.LeftNodeJoinCondition, nextSelect.Where)
+		nextSelect.Where = pgsql.OptionalAnd(traversalStep.EdgeConstraints.Expression, nextSelect.Where)
+		nextSelect.Where = pgsql.OptionalAnd(rightJoinExternal, nextSelect.Where)
+
+		return pgsql.Query{
+			Body: nextSelect,
+		}, nil
+	} else if traversalStep.RightNodeBound {
+		// Right node was already materialized in a previous frame.
+		//
+		// We have to promote that frame to the explicit JOIN root so that RightNodeJoinCondition can reference
+		// it in the ON clause. PostgreSQL forbids referencing a comma-joined table inside a subsequent
+		// explicit JOIN's ON clause.
+		leftJoinLocal, leftJoinExternal := partitionConstraintByLocality(
+			traversalStep.LeftNodeConstraints,
+			pgsql.AsIdentifierSet(traversalStep.LeftNode.Identifier, traversalStep.Edge.Identifier),
+		)
+
+		nextSelect.From = append(nextSelect.From, pgsql.FromClause{
+			Source: pgsql.TableReference{
+				Name:    pgsql.CompoundIdentifier{pgsql.TableEdge},
+				Binding: models.OptionalValue(traversalStep.Edge.Identifier),
+			},
+			Joins: []pgsql.Join{{
+				Table: pgsql.TableReference{
+					Name:    pgsql.CompoundIdentifier{pgsql.TableNode},
+					Binding: models.OptionalValue(traversalStep.LeftNode.Identifier),
+				},
+				JoinOperator: pgsql.JoinOperator{
+					JoinType:   pgsql.JoinTypeInner,
+					Constraint: pgsql.OptionalAnd(leftJoinLocal, traversalStep.LeftNodeJoinCondition),
+				},
+			}},
+		})
+
+		nextSelect.Where = pgsql.OptionalAnd(rightJoinLocal, nextSelect.Where)
+		nextSelect.Where = pgsql.OptionalAnd(traversalStep.RightNodeJoinCondition, nextSelect.Where)
+		nextSelect.Where = pgsql.OptionalAnd(leftJoinExternal, nextSelect.Where)
+		nextSelect.Where = pgsql.OptionalAnd(traversalStep.EdgeConstraints.Expression, nextSelect.Where)
+		nextSelect.Where = pgsql.OptionalAnd(rightJoinExternal, nextSelect.Where)
+
+		return pgsql.Query{
+			Body: nextSelect,
+		}, nil
+	} else {
+		// There is nothing to do to preserve outer bounds correlation - do the unbound traversal step
+		return s.buildTraversalPatternRoot(partFrame, traversalStep)
+	}
+}
+
 func (s *Translator) buildTraversalPatternRoot(partFrame *Frame, traversalStep *TraversalStep) (pgsql.Query, error) {
 	if traversalStep.Direction == graph.DirectionBoth {
 		return s.buildDirectionlessTraversalPatternRoot(traversalStep)

integration/testdata/bed6695.json

@@ -0,0 +1,114 @@
+{
+  "graph": {
+    "nodes": [
+      {
+        "id": "key_admins_empty",
+        "kinds": [
+          "NodeKind1"
+        ],
+        "properties": {
+          "name": "BED-6695 KEY ADMINS EMPTY"
+        }
+      },
+      {
+        "id": "key_admins_membered",
+        "kinds": [
+          "NodeKind1"
+        ],
+        "properties": {
+          "name": "BED-6695 KEY ADMINS MEMBERED"
+        }
+      },
+      {
+        "id": "member_user",
+        "kinds": [
+          "NodeKind2"
+        ],
+        "properties": {
+          "name": "BED-6695 MEMBER USER"
+        }
+      },
+      {
+        "id": "u1",
+        "kinds": [
+          "NodeKind1"
+        ],
+        "properties": {
+          "name": "User A"
+        }
+      },
+      {
+        "id": "u2",
+        "kinds": [
+          "NodeKind1"
+        ],
+        "properties": {
+          "name": "User B"
+        }
+      },
+      {
+        "id": "g1",
+        "kinds": [
+          "NodeKind2"
+        ],
+        "properties": {
+          "name": "KEY ADMINS ALPHA"
+        }
+      },
+      {
+        "id": "g2",
+        "kinds": [
+          "NodeKind2"
+        ],
+        "properties": {
+          "name": "KEY ADMINS BETA"
+        }
+      },
+      {
+        "id": "g3",
+        "kinds": [
+          "NodeKind2"
+        ],
+        "properties": {
+          "name": "OPERATORS"
+        }
+      },
+      {
+        "id": "g4",
+        "kinds": [
+          "NodeKind2"
+        ],
+        "properties": {
+          "name": "KEY ADMINS GAMMA"
+        }
+      }
+    ],
+    "edges": [
+      {
+        "start_id": "member_user",
+        "end_id": "key_admins_membered",
+        "kind": "EdgeKind1"
+      },
+      {
+        "start_id": "u1",
+        "end_id": "g2",
+        "kind": "EdgeKind1"
+      },
+      {
+        "start_id": "u2",
+        "end_id": "g3",
+        "kind": "EdgeKind1"
+      },
+      {
+        "start_id": "g3",
+        "end_id": "g1",
+        "kind": "EdgeKind1"
+      },
+      {
+        "start_id": "u1",
+        "end_id": "g1",
+        "kind": "EdgeKind2"
+      }
+    ]
+  }
+}