fix(query): expr children with projectset not fold count (#10827)

* fix(query): expr children with projectset not fold count

* if expr contains ProjectSet, not prune unused column

* ProjectSet precise_cardinality set None

* add count() from (unnest()) explain result

* fix ci: unnest explain result read bytes result diff

* modify explain test result diff

Author: TCeason

src/query/sql/src/planner/optimizer/heuristic/prune_unused_columns.rs

@@ -15,6 +15,7 @@
 use common_exception::ErrorCode;
 use common_exception::Result;
 
+use crate::optimizer::util::contaions_project_set;
 use crate::optimizer::ColumnSet;
 use crate::optimizer::SExpr;
 use crate::plans::Aggregate;
@@ -99,6 +100,14 @@ impl UnusedColumnPruner {
 
             RelOperator::EvalScalar(p) => {
                 let mut used = vec![];
+                if contaions_project_set(expr) {
+                    return Ok(SExpr::create_unary(
+                        RelOperator::EvalScalar(EvalScalar {
+                            items: p.items.clone(),
+                        }),
+                        expr.child(0)?.clone(),
+                    ));
+                }
                 // Only keep columns needed by parent plan.
                 for s in p.items.iter() {
                     if !required.contains(&s.index) {

src/query/sql/src/planner/optimizer/util.rs

@@ -13,6 +13,8 @@
 // limitations under the License.
 
 use super::SExpr;
+use crate::plans::Operator;
+use crate::plans::RelOp;
 use crate::plans::RelOperator;
 use crate::MetadataRef;
 
@@ -28,3 +30,16 @@ pub fn contains_local_table_scan(s_expr: &SExpr, metadata: &MetadataRef) -> bool
             false
         }
 }
+
+/// Check the expr contains ProjectSet op.
+pub fn contaions_project_set(s_expr: &SExpr) -> bool {
+    if let Some(child) = s_expr.children().iter().next() {
+        // Check children
+        return match child.plan.rel_op() {
+            RelOp::ProjectSet => true,
+            _ => contaions_project_set(child),
+        };
+    }
+
+    false
+}

src/query/sql/src/planner/plans/project_set.rs

@@ -57,6 +57,8 @@ impl Operator for ProjectSet {
             child_prop.output_columns.insert(srf.index);
         }
 
+        // ProjectSet is set-returning functions, precise_cardinality set None
+        child_prop.statistics.precise_cardinality = None;
         Ok(child_prop)
     }
 

tests/sqllogictests/suites/mode/standalone/explain/project_set.test

@@ -1,3 +1,48 @@
+statement ok
+use default;
+
+statement ok
+drop table if exists fold_count;
+
+statement ok
+create table fold_count(id array(int), c1 string);
+
+statement ok
+insert into fold_count values([1,2,3,4], 'x');
+
+query T
+explain select count() from (select unnest(id), c1 from fold_count)
+----
+EvalScalar
+├── expressions: [count() (#4)]
+├── estimated rows: 1.00
+└── AggregateFinal
+    ├── group by: []
+    ├── aggregate functions: [count()]
+    ├── estimated rows: 1.00
+    └── AggregatePartial
+        ├── group by: []
+        ├── aggregate functions: [count()]
+        ├── estimated rows: 1.00
+        └── EvalScalar
+            ├── expressions: [get(unnest (#2))]
+            ├── estimated rows: 1.00
+            └── ProjectSet
+                ├── estimated rows: 1.00
+                ├── set returning functions: unnest(CAST(fold_count.id (#0) AS Array(Int32 NULL)))
+                └── TableScan
+                    ├── table: default.default.fold_count
+                    ├── read rows: 1
+                    ├── read bytes: 87
+                    ├── partitions total: 1
+                    ├── partitions scanned: 1
+                    ├── pruning stats: [segments: <range pruning: 1 to 1>, blocks: <range pruning: 1 to 1, bloom pruning: 0 to 0>]
+                    ├── push downs: [filters: [], limit: NONE]
+                    └── estimated rows: 1.00
+
+statement ok
+drop table fold_count;
+
 statement ok
 drop database if exists project_set
 
@@ -10,32 +55,38 @@ use project_set
 query T
 explain select number from (select unnest([1,2,3]), number from numbers(10)) t
 ----
-ProjectSet
+EvalScalar
+├── expressions: [get(unnest (#1))]
 ├── estimated rows: 10.00
-├── set returning functions: unnest([1, 2, 3])
-└── TableScan
-    ├── table: default.system.numbers
-    ├── read rows: 10
-    ├── read bytes: 80
-    ├── partitions total: 1
-    ├── partitions scanned: 1
-    ├── push downs: [filters: [], limit: NONE]
-    └── estimated rows: 10.00
+   └── ProjectSet
+       ├── estimated rows: 10.00
+       ├── set returning functions: unnest([1, 2, 3])
+       └── TableScan
+           ├── table: default.system.numbers
+           ├── read rows: 10
+           ├── read bytes: 80
+           ├── partitions total: 1
+           ├── partitions scanned: 1
+           ├── push downs: [filters: [], limit: NONE]
+           └── estimated rows: 10.00
 
 query T
 explain select number from (select unnest([1,2,3,number]), number from numbers(10)) t
 ----
-ProjectSet
+EvalScalar
+├── expressions: [get(unnest (#1))]
 ├── estimated rows: 10.00
-├── set returning functions: unnest(CAST(array(1, 2, 3, numbers.number (#0)) AS Array(UInt64 NULL)))
-└── TableScan
-    ├── table: default.system.numbers
-    ├── read rows: 10
-    ├── read bytes: 80
-    ├── partitions total: 1
-    ├── partitions scanned: 1
-    ├── push downs: [filters: [], limit: NONE]
-    └── estimated rows: 10.00
+└── ProjectSet
+    ├── estimated rows: 10.00
+    ├── set returning functions: unnest(CAST(array(1, 2, 3, numbers.number (#0)) AS Array(UInt64 NULL)))
+    └── TableScan
+        ├── table: default.system.numbers
+        ├── read rows: 10
+        ├── read bytes: 80
+        ├── partitions total: 1
+        ├── partitions scanned: 1
+        ├── push downs: [filters: [], limit: NONE]
+        └── estimated rows: 10.00
 
 statement ok
 drop database project_set

tests/sqllogictests/suites/query/02_function/02_0062_function_unnest

@@ -369,6 +369,33 @@ drop table t;
 statement ok
 drop table t2;
 
+statement ok
+drop table if exists fold_count;
+
+statement ok
+create table fold_count(id array(int), c1 string);
+
+statement ok
+insert into fold_count values([1,2,3,4], 'x')
+
+query T
+select concat(c,'y') as b from (select concat(c::String, c1)  as c from (select unnest(id) as c, c1 from fold_count));
+----
+1xy
+2xy
+3xy
+4xy
+
+query I
+select count(b) from (select concat(c,'y') as b from (select concat(c::String, c1)  as c from (select unnest(id) as c, c1 from fold_count)));
+----
+4
+
+query I
+select count() from (select unnest(id), c1 from fold_count)
+----
+4
+
 statement ok
 DROP DATABASE db_02_0062;