fix(schema): require groupBy when having is present

CHANGELOG.md

@@ -7,6 +7,15 @@ Versioning follows Semantic Versioning with preview suffix `major.minor.patch-pr
 
 ---
 
+## [Unreleased]
+
+### ⚠️ Breaking Changes
+
+- **`having` now requires `groupBy`**: Queries that filter with `having` must also group their rows with `groupBy`. Previously the schema accepted `having` on its own, even though it has no meaning without groups. (#40)
+- **`groupBy` can no longer be empty**: `groupBy` must list at least one field. To skip grouping, leave the clause out.
+
+---
+
 ## [0.1.0-preview.0.5.0] - 2026-05-25
 
 Introduces sort direction control for `orderBy`. Previously `orderBy` accepted bare field references, making sort direction implementation-defined. Queries that used bare fields in `orderBy` must be migrated to the new `orderByItem` wrapper.

PureQL-Specification.json

@@ -3238,6 +3238,7 @@
     },
     "groupBy": {
       "type": "array",
+      "minItems": 1,
       "items": {
         "$ref": "#/definitions/field"
       }
@@ -3259,6 +3260,11 @@
     "from",
     "select"
   ],
+  "dependentRequired": {
+    "having": [
+      "groupBy"
+    ]
+  },
   "title": "PureQL specification",
   "type": "object"
 }

README.md

@@ -10,8 +10,8 @@ PureQL is a JSON-based declarative query language for relational data. Queries a
 | `select`   | yes      | Array of expressions to return |
 | `where`    | no       | Boolean filter applied before grouping |
 | `joins`    | no       | Array of join clauses |
-| `groupBy`  | no       | Fields to group rows by |
-| `having`   | no       | Boolean filter applied after grouping |
+| `groupBy`  | no       | Fields to group rows by (at least one) |
+| `having`   | no       | Boolean filter applied after grouping; requires `groupBy` |
 | `orderBy`  | no       | Fields to order results by |
 | `pagination` | no     | `skip` and `take` for paging |
 | `distinct` | no       | When `true`, deduplicate result rows (default: `false`) |
@@ -104,7 +104,7 @@ Each item in `select` is a value-returning expression (field, scalar, aggregate,
 They accept different shapes because they evaluate in different scopes:
 
 - **`where`** is evaluated per row. It accepts either a **boolean-returning** expression (single boolean) or a **boolean-array-returning** expression (per-row boolean column). Use the `each*` family for per-row predicates against fields.
-- **`having`** is evaluated per group, after `groupBy`. It accepts only a **boolean-returning** expression. Operands must reduce to a single value per group — typically aggregates compared with `greaterThan` / `equal` / etc. Per-row `each*` operators do **not** fit in `having`.
+- **`having`** is evaluated per group, after `groupBy`, and requires a non-empty `groupBy` to be present. It accepts only a **boolean-returning** expression. Operands must reduce to a single value per group — typically aggregates compared with `greaterThan` / `equal` / etc. Per-row `each*` operators do **not** fit in `having`.
 
 `where` example using per-row predicates: