feat(cubesql): PowerBI support for wrapped queries (#4752)

* feat(cubesql): PowerBI support for wrapped queries

* chore(cubesql): PowerBI support for wrapped queries -- support filters for aliased fields

* chore(cubesql): PowerBI support for wrapped queries -- replace column alias replacer with member pushdown replacer for centralized aliasing

* chore(cubesql): PowerBI support for wrapped queries -- remove unused replacers

* chore(cubesql): PowerBI support for wrapped queries -- list concat replacer for flattening old members search

Author: paveltiunov

Diff for: rust/cubesql/cubesql/src/compile/engine/df/scan.rs

@@ -1630,7 +1630,7 @@ impl QueryPlanner {
                     schema
                         .fields()
                         .iter()
-                        .map(|f| f.name().to_string())
+                        .map(|f| Some(f.name().to_string()))
                         .collect(),
                     query.request,
                     // @todo Remove after split!
@@ -3597,7 +3597,13 @@ mod tests {
                 order: None,
                 limit: None,
                 offset: None,
-                filters: None,
+                filters: Some(vec![V1LoadRequestQueryFilterItem {
+                    member: Some("KibanaSampleDataEcommerce.count".to_string()),
+                    operator: Some("gt".to_string()),
+                    values: Some(vec!["0".to_string()]),
+                    or: None,
+                    and: None,
+                }]),
             }
         );
 
@@ -3617,13 +3623,22 @@ mod tests {
                 order: None,
                 limit: None,
                 offset: None,
-                filters: Some(vec![V1LoadRequestQueryFilterItem {
-                    member: Some("KibanaSampleDataEcommerce.has_subscription".to_string()),
-                    operator: Some("equals".to_string()),
-                    values: Some(vec!["false".to_string()]),
-                    or: None,
-                    and: None,
-                }]),
+                filters: Some(vec![
+                    V1LoadRequestQueryFilterItem {
+                        member: Some("KibanaSampleDataEcommerce.has_subscription".to_string()),
+                        operator: Some("equals".to_string()),
+                        values: Some(vec!["false".to_string()]),
+                        or: None,
+                        and: None,
+                    },
+                    V1LoadRequestQueryFilterItem {
+                        member: Some("KibanaSampleDataEcommerce.count".to_string()),
+                        operator: Some("gt".to_string()),
+                        values: Some(vec!["0".to_string()]),
+                        or: None,
+                        and: None,
+                    }
+                ]),
             }
         );
     }
@@ -3884,7 +3899,13 @@ ORDER BY \"COUNT(count)\" DESC"
                 order: None,
                 limit: None,
                 offset: None,
-                filters: None,
+                filters: Some(vec![V1LoadRequestQueryFilterItem {
+                    member: Some("KibanaSampleDataEcommerce.count".to_string()),
+                    operator: Some("gt".to_string()),
+                    values: Some(vec!["0".to_string()]),
+                    or: None,
+                    and: None,
+                }]),
             }
         );
 
@@ -3895,17 +3916,11 @@ ORDER BY \"COUNT(count)\" DESC"
                 .iter()
                 .map(|f| f.name().to_string())
                 .collect::<Vec<_>>(),
-            vec![
-                "SUM(KibanaSampleDataEcommerce.count)".to_string(),
-                "COUNT(UInt8(1))".to_string()
-            ]
+            vec!["sum:count:ok".to_string(),]
         );
         assert_eq!(
             &cube_scan.member_fields,
-            &vec![
-                "KibanaSampleDataEcommerce.count".to_string(),
-                "KibanaSampleDataEcommerce.count".to_string()
-            ]
+            &vec![Some("KibanaSampleDataEcommerce.count".to_string())]
         );
     }
 
@@ -4113,7 +4128,7 @@ ORDER BY \"COUNT(count)\" DESC"
 
         let query_plan = convert_select_to_query_plan(
             "select \"rows\".\"customer_gender\" as \"customer_gender\",
-\n    count(1) as \"a0\"\
+\n    sum(\"rows\".\"count\") as \"a0\"\
 \nfrom\
 \n(\
 \n    select \"_\".\"count\",\
@@ -4140,7 +4155,7 @@ ORDER BY \"COUNT(count)\" DESC"
                 segments: Some(vec![]),
                 time_dimensions: None,
                 order: None,
-                limit: Some(1000001),
+                limit: Some(50000),
                 offset: None,
                 filters: Some(vec![V1LoadRequestQueryFilterItem {
                     member: Some("KibanaSampleDataEcommerce.customer_gender".to_string()),
@@ -4153,6 +4168,108 @@ ORDER BY \"COUNT(count)\" DESC"
         );
     }
 
+    #[test]
+    fn powerbi_inner_wrapped_dates() {
+        init_logger();
+
+        let query_plan = convert_select_to_query_plan(
+            "select \"_\".\"created_at_day\",\
+\n    \"_\".\"a0\"\
+\nfrom \
+\n(\
+\n    select \"rows\".\"created_at_day\" as \"created_at_day\",\
+\n        sum(\"rows\".\"cnt\") as \"a0\"\
+\n    from \
+\n    (\
+\n        select count(*) cnt,date_trunc('day', order_date) as created_at_day, date_trunc('month', order_date) as created_at_month from public.KibanaSampleDataEcommerce group by 2, 3\
+\n    ) \"rows\"\
+\n    group by \"created_at_day\"\
+\n) \"_\"\
+\nwhere not \"_\".\"a0\" is null\
+\nlimit 1000001"
+                .to_string(),
+            DatabaseProtocol::PostgreSQL,
+        );
+
+        let logical_plan = query_plan.as_logical_plan();
+        assert_eq!(
+            logical_plan.find_cube_scan().request,
+            V1LoadRequestQuery {
+                measures: Some(vec!["KibanaSampleDataEcommerce.count".to_string()]),
+                dimensions: Some(vec![]),
+                segments: Some(vec![]),
+                time_dimensions: Some(vec![V1LoadRequestQueryTimeDimension {
+                    dimension: "KibanaSampleDataEcommerce.order_date".to_string(),
+                    granularity: Some("day".to_string()),
+                    date_range: None,
+                }]),
+                order: None,
+                limit: Some(50000),
+                offset: None,
+                filters: Some(vec![V1LoadRequestQueryFilterItem {
+                    member: Some("KibanaSampleDataEcommerce.count".to_string()),
+                    operator: Some("set".to_string()),
+                    values: None,
+                    or: None,
+                    and: None,
+                }]),
+            }
+        );
+    }
+
+    #[test]
+    fn powerbi_inner_wrapped_asterisk() {
+        init_logger();
+
+        let query_plan = convert_select_to_query_plan(
+            "select \"rows\".\"customer_gender\" as \"customer_gender\",\
+\n    \"rows\".\"created_at_month\" as \"created_at_month\"\
+\nfrom \
+\n(\
+\n    select \"_\".\"count\",\
+\n        \"_\".\"minPrice\",\
+\n        \"_\".\"maxPrice\",\
+\n        \"_\".\"avgPrice\",\
+\n        \"_\".\"order_date\",\
+\n        \"_\".\"customer_gender\",\
+\n        \"_\".\"created_at_day\",\
+\n        \"_\".\"created_at_month\"\
+\n    from \
+\n    (\
+\n        select *, date_trunc('day', order_date) created_at_day, date_trunc('month', order_date) created_at_month from public.KibanaSampleDataEcommerce\
+\n    ) \"_\"\
+\n    where \"_\".\"created_at_month\" < timestamp '2022-06-13 00:00:00' and \"_\".\"created_at_month\" >= timestamp '2021-12-16 00:00:00'\
+\n) \"rows\"\
+\ngroup by \"customer_gender\",\
+\n    \"created_at_month\"\
+\nlimit 1000001"
+                .to_string(),
+            DatabaseProtocol::PostgreSQL,
+        );
+
+        let logical_plan = query_plan.as_logical_plan();
+        assert_eq!(
+            logical_plan.find_cube_scan().request,
+            V1LoadRequestQuery {
+                measures: Some(vec![]),
+                dimensions: Some(vec!["KibanaSampleDataEcommerce.customer_gender".to_string()]),
+                segments: Some(vec![]),
+                time_dimensions: Some(vec![V1LoadRequestQueryTimeDimension {
+                    dimension: "KibanaSampleDataEcommerce.order_date".to_string(),
+                    granularity: Some("month".to_string()),
+                    date_range: Some(json!(vec![
+                        "2021-12-16 00:00:00".to_string(),
+                        "2022-06-13 00:00:00".to_string()
+                    ])),
+                }]),
+                order: None,
+                limit: Some(50000),
+                offset: None,
+                filters: None,
+            }
+        );
+    }
+
     #[test]
     fn test_select_aggregations() {
         let variants = vec![
@@ -4304,10 +4421,10 @@ ORDER BY \"COUNT(count)\" DESC"
             //     "SELECT is_male FROM KibanaSampleDataEcommerce".to_string(),
             //     CompilationError::User("Unable to use segment 'is_male' as column in SELECT statement".to_string()),
             // ),
-            (
-                "SELECT COUNT(*) FROM KibanaSampleDataEcommerce GROUP BY is_male".to_string(),
-                CompilationError::User("Unable to use segment 'is_male' in GROUP BY".to_string()),
-            ),
+            // (
+            //     "SELECT COUNT(*) FROM KibanaSampleDataEcommerce GROUP BY is_male".to_string(),
+            //     CompilationError::User("Unable to use segment 'is_male' in GROUP BY".to_string()),
+            // ),
             // (
             //     "SELECT COUNT(*) FROM KibanaSampleDataEcommerce ORDER BY is_male DESC".to_string(),
             //     CompilationError::User("Unable to use segment 'is_male' in ORDER BY".to_string()),
@@ -4617,7 +4734,6 @@ ORDER BY \"COUNT(count)\" DESC"
                     ""
                 }
             );
-            println!("Query: {}", query);
             let logical_plan =
                 convert_select_to_query_plan(query, DatabaseProtocol::MySQL).as_logical_plan();
 

Diff for: rust/cubesql/cubesql/src/compile/mod.rs

@@ -4,7 +4,7 @@ use crate::{
         rewrite::{
             converter::{is_expr_node, node_to_expr},
             AliasExprAlias, ColumnExprColumn, DimensionName, LiteralExprValue, LogicalPlanLanguage,
-            MeasureName, TableScanSourceTableName, TimeDimensionName,
+            MeasureName, SegmentName, TableScanSourceTableName, TimeDimensionName,
         },
     },
     var_iter, CubeError,
@@ -124,6 +124,16 @@ impl LogicalPlanAnalysis {
                     None
                 }
             }
+            LogicalPlanLanguage::Segment(params) => {
+                if let Some(_) = column_name(params[1]) {
+                    let expr = original_expr(params[1])?;
+                    let segment_name = var_iter!(egraph[params[0]], SegmentName).next().unwrap();
+                    map.push((segment_name.to_string(), expr));
+                    Some(map)
+                } else {
+                    None
+                }
+            }
             LogicalPlanLanguage::TimeDimension(params) => {
                 if let Some(_) = column_name(params[3]) {
                     let expr = original_expr(params[3])?;