chore(tesseract): Fix for multi stage with sub queries

Author: waralexrom

packages/cubejs-schema-compiler/test/integration/postgres/multi-stage.test.ts

@@ -0,0 +1,201 @@
+import {
+  getEnv,
+} from '@cubejs-backend/shared';
+import { PostgresQuery } from '../../../src/adapter/PostgresQuery';
+import { prepareYamlCompiler } from '../../unit/PrepareCompiler';
+import { dbRunner } from './PostgresDBRunner';
+
+describe('Multi-Stage', () => {
+  jest.setTimeout(200000);
+
+  const { compiler, joinGraph, cubeEvaluator } = prepareYamlCompiler(`
+cubes:
+  - name: orders
+    sql: >
+      SELECT 10 as ID, 'complited' as STATUS, '2024-01-12T20:00:00.000Z'::timestamptz as CREATED_AT
+    joins:
+      - name: line_items
+        sql: "{CUBE}.ID = {line_items}.order_id"
+        relationship: many_to_one
+
+    dimensions:
+      - name: id
+        sql: ID
+        type: number
+        primary_key: true
+
+      - name: status
+        sql: STATUS
+        type: string
+
+      - name: date
+        sql: CREATED_AT
+        type: time
+
+      - name: amount
+        sql: '{line_items.total_amount}'
+        type: number
+        sub_query: true
+
+    measures:
+      - name: count
+        type: count
+
+      - name: completed_count
+        type: count
+        filters:
+          - sql: "{CUBE}.STATUS = 'completed'"
+
+      - name: returned_count
+        type: count
+        filters:
+          - sql: "{CUBE}.STATUS = 'returned'"
+
+      - name: return_rate
+        type: number
+        sql: "({returned_count} / NULLIF({completed_count}, 0)) * 100.0"
+        description: "Percentage of returned orders out of completed, exclude just placed orders."
+        format: percent
+
+      - name: total_amount
+        sql: '{CUBE.amount}'
+        type: sum
+
+      - name: revenue
+        sql: "CASE WHEN {CUBE}.status = 'completed' THEN {CUBE.amount} END"
+        type: sum
+        format: currency
+
+      - name: average_order_value
+        sql: '{CUBE.amount}'
+        type: avg
+
+      - name: revenue_1_y_ago
+        sql: "{revenue}"
+        multi_stage: true
+        type: number
+        format: currency
+        time_shift:
+          - time_dimension: date
+            interval: 1 year
+            type: prior
+          - time_dimension: orders_view.date
+            interval: 1 year
+            type: prior
+
+      - name: cagr_1_y
+        sql: "(({revenue} / {revenue_1_y_ago}) - 1)"
+        type: number
+        format: percent
+        description: "Annual CAGR, year over year growth in revenue"
+
+  - name: line_items
+    sql: >
+      SELECT 10 as ID, '2024-01-12T20:00:00.000Z'::timestamptz as CREATED_AT, 10 as ORDER_ID, 10 as PRODUCT_ID
+    public: false
+
+    joins:
+      - name: products
+        sql: "{CUBE}.PRODUCT_ID = {products}.ID"
+        relationship: many_to_one
+
+    dimensions:
+      - name: id
+        sql: ID
+        type: number
+        primary_key: true
+
+      - name: created_at
+        sql: CREATED_AT
+        type: time
+
+      - name: price
+        sql: "{products.price}"
+        type: number
+
+    measures:
+      - name: count
+        type: count
+
+      - name: total_amount
+        sql: "{price}"
+        type: sum
+
+  - name: products
+    sql: >
+      SELECT 10 as ID, 'some category' as PRODUCT_CATEGORY, 'some name' as NAME, 10 as PRICE
+    public: false
+    description: >
+      Products and categories in our e-commerce store.
+
+    dimensions:
+      - name: id
+        sql: ID
+        type: number
+        primary_key: true
+
+      - name: product_category
+        sql: PRODUCT_CATEGORY
+        type: string
+
+      - name: name
+        sql: NAME
+        type: string
+
+      - name: price
+        sql: PRICE
+        type: number
+
+    measures:
+      - name: count
+        type: count
+views:
+  - name: orders_view
+
+    cubes:
+      - join_path: orders
+        includes:
+          - date
+          - revenue
+          - cagr_1_y
+          - return_rate
+
+      - join_path: line_items.products
+        prefix: true
+        includes:
+          - product_category
+
+    `);
+
+  async function runQueryTest(q, expectedResult) {
+    if (!getEnv('nativeSqlPlanner')) {
+      return;
+    }
+    await compiler.compile();
+    const query = new PostgresQuery({ joinGraph, cubeEvaluator, compiler }, q);
+
+    console.log(query.buildSqlAndParams());
+
+    const res = await dbRunner.testQuery(query.buildSqlAndParams());
+    console.log(JSON.stringify(res));
+
+    expect(res).toEqual(
+      expectedResult
+    );
+  }
+
+  it('multi stage over sub query', async () => runQueryTest({
+    measures: ['orders.revenue', 'orders.revenue_1_y_ago'],
+    timeDimensions: [
+      {
+        dimension: 'orders.date',
+        granularity: 'year'
+      }
+    ],
+    dimensions: [
+      'orders.status',
+      'orders.amount'
+    ],
+    timezone: 'UTC'
+  }, []));
+});

rust/cubesqlplanner/cubesqlplanner/src/planner/sql_evaluator/collectors/sub_query_dimensions.rs

@@ -20,6 +20,9 @@ impl SubQueryDimensionsCollector {
 
     pub fn extract_result(self) -> Vec<Rc<MemberSymbol>> {
         self.sub_query_dimensions
+            .into_iter()
+            .unique_by(|m| m.full_name())
+            .collect()
     }
 
     fn check_dim_has_measures(&self, dim: &DimensionSymbol) -> bool {