planner: donot prune all columns for Projection (pingcap#24024)

cmd/explaintest/r/explain_easy.result

@@ -194,29 +194,29 @@ test	t4	1	expr_idx	1	NULL	NULL		(`a` + `b` + 1)	2	YES	NO
 explain format = 'brief' select count(1) from (select count(1) from (select * from t1 where c3 = 100) k) k2;
 id	estRows	task	access object	operator info
 StreamAgg	1.00	root		funcs:count(1)->Column#5
-└─StreamAgg	1.00	root		funcs:firstrow(Column#9)->Column#7
+└─StreamAgg	1.00	root		funcs:firstrow(Column#13)->Column#11
   └─TableReader	1.00	root		data:StreamAgg
-    └─StreamAgg	1.00	cop[tikv]		funcs:firstrow(1)->Column#9
+    └─StreamAgg	1.00	cop[tikv]		funcs:firstrow(1)->Column#13
       └─Selection	10.00	cop[tikv]		eq(test.t1.c3, 100)
         └─TableFullScan	10000.00	cop[tikv]	table:t1	keep order:false, stats:pseudo
 explain format = 'brief' select 1 from (select count(c2), count(c3) from t1) k;
 id	estRows	task	access object	operator info
 Projection	1.00	root		1->Column#6
-└─StreamAgg	1.00	root		funcs:firstrow(Column#14)->Column#9
+└─StreamAgg	1.00	root		funcs:firstrow(Column#16)->Column#11
   └─TableReader	1.00	root		data:StreamAgg
-    └─StreamAgg	1.00	cop[tikv]		funcs:firstrow(1)->Column#14
+    └─StreamAgg	1.00	cop[tikv]		funcs:firstrow(1)->Column#16
       └─TableFullScan	10000.00	cop[tikv]	table:t1	keep order:false, stats:pseudo
 explain format = 'brief' select count(1) from (select max(c2), count(c3) as m from t1) k;
 id	estRows	task	access object	operator info
 StreamAgg	1.00	root		funcs:count(1)->Column#6
-└─StreamAgg	1.00	root		funcs:firstrow(Column#13)->Column#8
+└─StreamAgg	1.00	root		funcs:firstrow(Column#15)->Column#10
   └─TableReader	1.00	root		data:StreamAgg
-    └─StreamAgg	1.00	cop[tikv]		funcs:firstrow(1)->Column#13
+    └─StreamAgg	1.00	cop[tikv]		funcs:firstrow(1)->Column#15
       └─TableFullScan	10000.00	cop[tikv]	table:t1	keep order:false, stats:pseudo
 explain format = 'brief' select count(1) from (select count(c2) from t1 group by c3) k;
 id	estRows	task	access object	operator info
 StreamAgg	1.00	root		funcs:count(1)->Column#5
-└─HashAgg	8000.00	root		group by:test.t1.c3, funcs:firstrow(1)->Column#7
+└─HashAgg	8000.00	root		group by:test.t1.c3, funcs:firstrow(1)->Column#9
   └─TableReader	10000.00	root		data:TableFullScan
     └─TableFullScan	10000.00	cop[tikv]	table:t1	keep order:false, stats:pseudo
 set @@session.tidb_opt_insubq_to_join_and_agg=0;
@@ -498,8 +498,8 @@ PRIMARY KEY (`id`)
 explain format = 'brief' SELECT COUNT(1) FROM (SELECT COALESCE(b.region_name, '不详') region_name, SUM(a.registration_num) registration_num FROM (SELECT stat_date, show_date, region_id, 0 registration_num FROM test01 WHERE period = 1 AND stat_date >= 20191202 AND stat_date <= 20191202 UNION ALL SELECT stat_date, show_date, region_id, registration_num registration_num FROM test01 WHERE period = 1 AND stat_date >= 20191202 AND stat_date <= 20191202) a LEFT JOIN test02 b ON a.region_id = b.id WHERE registration_num > 0 AND a.stat_date >= '20191202' AND a.stat_date <= '20191202' GROUP BY a.stat_date , a.show_date , COALESCE(b.region_name, '不详') ) JLS;
 id	estRows	task	access object	operator info
 StreamAgg	1.00	root		funcs:count(1)->Column#22
-└─HashAgg	8000.00	root		group by:Column#32, Column#33, Column#34, funcs:firstrow(1)->Column#31
-  └─Projection	10000.01	root		Column#14, Column#15, coalesce(test.test02.region_name, 不详)->Column#34
+└─HashAgg	8000.00	root		group by:Column#34, Column#35, Column#36, funcs:firstrow(1)->Column#33
+  └─Projection	10000.01	root		Column#14, Column#15, coalesce(test.test02.region_name, 不详)->Column#36
     └─HashJoin	10000.01	root		left outer join, equal:[eq(Column#16, test.test02.id)]
       ├─TableReader(Build)	10000.00	root		data:TableFullScan
       │ └─TableFullScan	10000.00	cop[tikv]	table:b	keep order:false, stats:pseudo

cmd/explaintest/r/explain_join_stats.result

@@ -7,13 +7,14 @@ load stats 's/explain_join_stats_lo.json';
 explain format = 'brief' select count(*) from e, lo where lo.a=e.a and e.b=22336;
 id	estRows	task	access object	operator info
 StreamAgg	1.00	root		funcs:count(1)->Column#5
-└─HashJoin	19977.00	root		inner join, equal:[eq(test.lo.a, test.e.a)]
-  ├─TableReader(Build)	250.00	root		data:TableFullScan
-  │ └─TableFullScan	250.00	cop[tikv]	table:lo	keep order:false
-  └─IndexLookUp(Probe)	19977.00	root		
-    ├─IndexRangeScan(Build)	19977.00	cop[tikv]	table:e, index:idx_b(b)	range:[22336,22336], keep order:false
-    └─Selection(Probe)	19977.00	cop[tikv]		not(isnull(test.e.a))
-      └─TableRowIDScan	19977.00	cop[tikv]	table:e	keep order:false
+└─Projection	19977.00	root		1->Column#6
+  └─HashJoin	19977.00	root		inner join, equal:[eq(test.lo.a, test.e.a)]
+    ├─TableReader(Build)	250.00	root		data:TableFullScan
+    │ └─TableFullScan	250.00	cop[tikv]	table:lo	keep order:false
+    └─IndexLookUp(Probe)	19977.00	root		
+      ├─IndexRangeScan(Build)	19977.00	cop[tikv]	table:e, index:idx_b(b)	range:[22336,22336], keep order:false
+      └─Selection(Probe)	19977.00	cop[tikv]		not(isnull(test.e.a))
+        └─TableRowIDScan	19977.00	cop[tikv]	table:e	keep order:false
 explain format = 'brief' select /*+ TIDB_INLJ(e) */ count(*) from e, lo where lo.a=e.a and e.b=22336;
 id	estRows	task	access object	operator info
 StreamAgg	1.00	root		funcs:count(1)->Column#5

planner/core/integration_test.go

@@ -3110,6 +3110,31 @@ func (s *testIntegrationSerialSuite) TestLimitIndexLookUpKeepOrder(c *C) {
 	}
 }
 
+// Apply operator may got panic because empty Projection is eliminated.
+func (s *testIntegrationSerialSuite) TestIssue23887(c *C) {
+	tk := testkit.NewTestKit(c, s.store)
+	tk.MustExec("use test")
+	tk.MustExec("drop table if exists t;")
+	tk.MustExec("create table t(a int, b int);")
+	tk.MustExec("insert into t values(1, 2), (3, 4);")
+	var input []string
+	var output []struct {
+		SQL  string
+		Plan []string
+		Res  []string
+	}
+	s.testData.GetTestCases(c, &input, &output)
+	for i, tt := range input {
+		s.testData.OnRecord(func() {
+			output[i].SQL = tt
+			output[i].Plan = s.testData.ConvertRowsToStrings(tk.MustQuery("explain format = 'brief' " + tt).Rows())
+			output[i].Res = s.testData.ConvertRowsToStrings(tk.MustQuery(tt).Sort().Rows())
+		})
+		tk.MustQuery("explain format = 'brief' " + tt).Check(testkit.Rows(output[i].Plan...))
+		tk.MustQuery(tt).Sort().Check(testkit.Rows(output[i].Res...))
+	}
+}
+
 func (s *testIntegrationSuite) TestDecorrelateInnerJoinInSubquery(c *C) {
 	tk := testkit.NewTestKit(c, s.store)
 

planner/core/rule_column_pruning.go

@@ -71,6 +71,18 @@ func (p *LogicalProjection) PruneColumns(parentUsedCols []*expression.Column) er
 			p.Exprs = append(p.Exprs[:i], p.Exprs[i+1:]...)
 		}
 	}
+	// Here we add a constant 1 to avoid Projection operator is eliminated. (#23887)
+	if len(p.Exprs) == 0 {
+		constOne := expression.NewOne()
+		p.schema.Append(&expression.Column{
+			UniqueID: p.ctx.GetSessionVars().AllocPlanColumnID(),
+			RetType:  constOne.GetType(),
+		})
+		p.Exprs = append(p.Exprs, &expression.Constant{
+			Value:   constOne.Value,
+			RetType: constOne.GetType(),
+		})
+	}
 	selfUsedCols := make([]*expression.Column, 0, len(p.Exprs))
 	selfUsedCols = expression.ExtractColumnsFromExpressions(selfUsedCols, p.Exprs, nil)
 	return child.PruneColumns(selfUsedCols)

planner/core/rule_eliminate_projection.go

@@ -60,9 +60,6 @@ func canProjectionBeEliminatedStrict(p *PhysicalProjection) bool {
 	if p.CalculateNoDelay {
 		return false
 	}
-	if p.Schema().Len() == 0 {
-		return true
-	}
 	child := p.Children()[0]
 	if p.Schema().Len() != child.Schema().Len() {
 		return false

planner/core/testdata/integration_serial_suite_in.json

@@ -281,5 +281,11 @@
       "desc format = 'brief' select * from t where a = 1 and b > 2 and b < 10 and d = 10 order by b,c limit 10",
       "desc format = 'brief' select * from t where a = 1 and b > 2 and b < 10 and d = 10 order by b desc, c desc limit 10"
     ]
+  },
+  {
+    "name": "TestIssue23887",
+    "cases": [
+        "select (2) in (select b from t) from (select t.a < (select t.a from t t1 limit 1) from t) t"
+    ]
   }
 ]