statistics: change default value of tidb_analyze_version to 1 (#31826)

close #31748

Author: yiwen92

cmd/explaintest/r/subquery.result

@@ -22,9 +22,9 @@ Projection	5.00	root		Column#22
   ├─TableReader(Build)	5.00	root		data:TableFullScan
   │ └─TableFullScan	5.00	cop[tikv]	table:t	keep order:false
   └─StreamAgg(Probe)	1.00	root		funcs:count(1)->Column#21
-    └─IndexJoin	0.22	root		inner join, inner:TableReader, outer key:test.t.a, inner key:test.t.a, equal cond:eq(test.t.a, test.t.a)
-      ├─IndexReader(Build)	0.45	root		index:IndexRangeScan
-      │ └─IndexRangeScan	0.45	cop[tikv]	table:s, index:idx(b, c, d)	range: decided by [eq(test.t.b, 1) eq(test.t.c, 1) eq(test.t.d, test.t.a)], keep order:false
+    └─IndexJoin	0.50	root		inner join, inner:TableReader, outer key:test.t.a, inner key:test.t.a, equal cond:eq(test.t.a, test.t.a)
+      ├─IndexReader(Build)	1.00	root		index:IndexRangeScan
+      │ └─IndexRangeScan	1.00	cop[tikv]	table:s, index:idx(b, c, d)	range: decided by [eq(test.t.b, 1) eq(test.t.c, 1) eq(test.t.d, test.t.a)], keep order:false
       └─TableReader(Probe)	1.00	root		data:TableRangeScan
         └─TableRangeScan	1.00	cop[tikv]	table:t1	range: decided by [test.t.a], keep order:false
 drop table if exists t;

executor/analyze_test.go

@@ -413,8 +413,8 @@ func (s *testFastAnalyze) TestFastAnalyze(c *C) {
 	tk.MustExec("use test")
 	tk.MustExec("drop table if exists t")
 	tk.MustExec("create table t(a int primary key, b int, c char(10), index index_b(b))")
-	tk.MustExec("set @@session.tidb_enable_fast_analyze=1")
-	tk.MustExec("set @@session.tidb_build_stats_concurrency=1")
+	tk.MustExec("set @@session.tidb_enable_fast_analyze = 1")
+	tk.MustExec("set @@session.tidb_build_stats_concurrency = 1")
 	tk.MustExec("set @@tidb_analyze_version = 1")
 	// Should not panic.
 	tk.MustExec("analyze table t")
@@ -564,18 +564,18 @@ func (s *testSerialSuite2) TestAnalyzeIndex(c *C) {
 	tk.MustExec("drop table if exists t1")
 	tk.MustExec("create table t1 (id int, v int, primary key(id), index k(v))")
 	tk.MustExec("insert into t1(id, v) values(1, 2), (2, 2), (3, 2), (4, 2), (5, 1), (6, 3), (7, 4)")
-	tk.MustExec("set @@tidb_analyze_version=1")
+	tk.MustExec("set @@tidb_analyze_version = 1")
 	tk.MustExec("analyze table t1 index k")
 	c.Assert(len(tk.MustQuery("show stats_buckets where table_name = 't1' and column_name = 'k' and is_index = 1").Rows()), Greater, 0)
 	tk.MustExec("set @@tidb_analyze_version=default")
 	tk.MustExec("analyze table t1")
-	c.Assert(len(tk.MustQuery("show stats_topn where table_name = 't1' and column_name = 'k' and is_index = 1").Rows()), Greater, 0)
+	c.Assert(len(tk.MustQuery("show stats_topn where table_name = 't1' and column_name = 'k' and is_index = 1").Rows()), Equals, 0)
 
 	func() {
 		defer tk.MustExec("set @@session.tidb_enable_fast_analyze=0")
 		tk.MustExec("drop stats t1")
-		tk.MustExec("set @@session.tidb_enable_fast_analyze=1")
-		tk.MustExec("set @@tidb_analyze_version=1")
+		tk.MustExec("set @@session.tidb_enable_fast_analyze = 1")
+		tk.MustExec("set @@tidb_analyze_version = 1")
 		tk.MustExec("analyze table t1 index k")
 		c.Assert(len(tk.MustQuery("show stats_buckets where table_name = 't1' and column_name = 'k' and is_index = 1").Rows()), Greater, 1)
 	}()
@@ -863,7 +863,7 @@ func (s *testSuite1) TestNormalAnalyzeOnCommonHandle(c *C) {
 	tk.MustExec("insert into t3 values(1,1,1), (2,2,2), (3,3,3)")
 
 	// Version2 is tested in TestStatsVer2.
-	tk.MustExec("set@@tidb_analyze_version=1")
+	tk.MustExec("set@@tidb_analyze_version = 1")
 	tk.MustExec("analyze table t1, t2, t3")
 
 	tk.MustQuery(`show stats_buckets where table_name in ("t1", "t2", "t3")`).Sort().Check(testkit.Rows(
@@ -913,6 +913,8 @@ func (s *testSuite1) TestDefaultValForAnalyze(c *C) {
 	for i := 1; i < 4; i++ {
 		tk.MustExec("insert into t values (?)", i)
 	}
+	tk.MustQuery("select @@tidb_analyze_version").Check(testkit.Rows("1"))
+	tk.MustQuery("select @@session.tidb_analyze_version").Check(testkit.Rows("1"))
 	tk.MustQuery("select @@tidb_enable_fast_analyze").Check(testkit.Rows("0"))
 	tk.MustQuery("select @@session.tidb_enable_fast_analyze").Check(testkit.Rows("0"))
 	tk.MustExec("analyze table t with 0 topn;")
@@ -956,7 +958,7 @@ func (s *testSerialSuite2) TestIssue20874(c *C) {
 	tk.MustExec("delete from mysql.stats_histograms")
 	tk.MustExec("create table t (a char(10) collate utf8mb4_unicode_ci not null, b char(20) collate utf8mb4_general_ci not null, key idxa(a), key idxb(b))")
 	tk.MustExec("insert into t values ('#', 'C'), ('$', 'c'), ('a', 'a')")
-	tk.MustExec("set @@tidb_analyze_version=1")
+	tk.MustExec("set @@tidb_analyze_version = 1")
 	tk.MustExec("analyze table t")
 	tk.MustQuery("show stats_buckets where db_name = 'test' and table_name = 't'").Sort().Check(testkit.Rows(
 		"test t  a 0 0 1 1 \x02\xd2 \x02\xd2 0",

planner/core/testdata/analyze_suite_out.json

@@ -435,14 +435,14 @@
   {
     "Name": "TestAnalyze",
     "Cases": [
-      "Analyze{Table(_tidb_rowid, a, b, _tidb_rowid)}",
+      "Analyze{Index(a),Table(a, b)}",
       "TableReader(Table(t)->Sel([le(test.t.a, 2)]))",
       "IndexReader(Index(t.b)[[-inf,2)])",
       "TableReader(Table(t)->Sel([eq(test.t.a, 1) le(test.t.b, 2)]))",
       "TableReader(Table(t1)->Sel([le(test.t1.a, 2)]))",
       "IndexLookUp(Index(t1.a)[[1,1]], Table(t1)->Sel([le(test.t1.b, 2)]))",
       "TableReader(Table(t2)->Sel([le(test.t2.a, 2)]))",
-      "Analyze{Table(_tidb_rowid, a, b, _tidb_rowid)}",
+      "Analyze{Index(a),Index(b)}",
       "PartitionUnionAll{TableReader(Table(t4)->Sel([le(test.t4.a, 2)]))->TableReader(Table(t4)->Sel([le(test.t4.a, 2)]))}",
       "PartitionUnionAll{IndexReader(Index(t4.b)[[-inf,2)])->IndexReader(Index(t4.b)[[-inf,2)])}",
       "TableReader(Table(t4)->Sel([eq(test.t4.a, 1) le(test.t4.b, 2)]))"

planner/core/testdata/integration_suite_out.json

@@ -984,8 +984,8 @@
       {
         "SQL": "select * from t1 where t1.a = 1 and t1.b < \"333\"",
         "Plan": [
-          "TableReader 0.82 root  data:TableRangeScan",
-          "└─TableRangeScan 0.82 cop[tikv] table:t1 range:[1 -inf,1 \"333\"), keep order:false"
+          "TableReader 0.67 root  data:TableRangeScan",
+          "└─TableRangeScan 0.67 cop[tikv] table:t1 range:[1 -inf,1 \"333\"), keep order:false"
         ],
         "Res": [
           "1 111 1.1000000000 11"

sessionctx/variable/tidb_vars.go

@@ -715,7 +715,7 @@ const (
 	DefTiDBEnableAsyncCommit           = false
 	DefTiDBEnable1PC                   = false
 	DefTiDBGuaranteeLinearizability    = true
-	DefTiDBAnalyzeVersion              = 2
+	DefTiDBAnalyzeVersion              = 1
 	DefTiDBEnableIndexMergeJoin        = false
 	DefTiDBTrackAggregateMemoryUsage   = true
 	DefTiDBEnableExchangePartition     = false

statistics/handle/ddl_test.go

@@ -179,8 +179,6 @@ func (s *testStatsSuite) TestDDLHistogram(c *C) {
 	rs := testKit.MustQuery("select count(*) from mysql.stats_histograms where table_id = ? and hist_id = 1 and is_index =1", tableInfo.ID)
 	rs.Check(testkit.Rows("1"))
 	rs = testKit.MustQuery("select count(*) from mysql.stats_buckets where table_id = ? and hist_id = 1 and is_index = 1", tableInfo.ID)
-	rs.Check(testkit.Rows("0"))
-	rs = testKit.MustQuery("select count(*) from mysql.stats_top_n where table_id = ? and hist_id = 1 and is_index = 1", tableInfo.ID)
 	rs.Check(testkit.Rows("2"))
 }
 

statistics/handle/update_test.go

@@ -504,8 +504,7 @@ func (s *testStatsSuite) TestAutoUpdate(c *C) {
 		hg, ok := stats.Indices[tableInfo.Indices[0].ID]
 		c.Assert(ok, IsTrue)
 		c.Assert(hg.NDV, Equals, int64(3))
-		c.Assert(hg.Len(), Equals, 0)
-		c.Assert(hg.TopN.Num(), Equals, 3)
+		c.Assert(hg.Len(), Equals, 3)
 	})
 }
 

statistics/testdata/stats_suite_out.json

@@ -607,7 +607,7 @@
       },
       {
         "SQL": "select * from t where a < 8 and (b > 10 or c < 3 or b > 4) and a > 2",
-        "Selectivity": 0
+        "Selectivity": 0.3125
       }
     ]
   }

util/ranger/testdata/ranger_suite_out.json

@@ -265,8 +265,8 @@
       {
         "SQL": "select * from t where a = 1 and ((b = 1) or (b = 2 and c = 3));",
         "Plan": [
-          "TableReader_6 1.71 root  data:TableRangeScan_5",
-          "└─TableRangeScan_5 1.71 cop[tikv] table:t range:[1 1,1 1], [1 2 3,1 2 3], keep order:false"
+          "TableReader_6 2.00 root  data:TableRangeScan_5",
+          "└─TableRangeScan_5 2.00 cop[tikv] table:t range:[1 1,1 1], [1 2 3,1 2 3], keep order:false"
         ],
         "Result": [
           "1 1 1"
@@ -293,8 +293,8 @@
       {
         "SQL": "select * from t use index(primary) where ((a = 1) or (a = 2 and b = 2)) and c = 3;",
         "Plan": [
-          "TableReader_7 0.68 root  data:Selection_6",
-          "└─Selection_6 0.68 cop[tikv]  eq(test.t.c, 3), or(eq(test.t.a, 1), and(eq(test.t.a, 2), eq(test.t.b, 2)))",
+          "TableReader_7 0.75 root  data:Selection_6",
+          "└─Selection_6 0.75 cop[tikv]  eq(test.t.c, 3), or(eq(test.t.a, 1), and(eq(test.t.a, 2), eq(test.t.b, 2)))",
           "  └─TableRangeScan_5 2.00 cop[tikv] table:t range:[1,1], [2,2], keep order:false"
         ],
         "Result": [
@@ -304,7 +304,7 @@
       {
         "SQL": "select * from t where (a,b) in ((1,1),(2,2)) and c > 2 and (a,b,c) in ((1,1,1),(2,2,3));",
         "Plan": [
-          "Selection_6 0.44 root  gt(test.t.c, 2), or(and(eq(test.t.a, 1), eq(test.t.b, 1)), and(eq(test.t.a, 2), eq(test.t.b, 2)))",
+          "Selection_6 0.56 root  gt(test.t.c, 2), or(and(eq(test.t.a, 1), eq(test.t.b, 1)), and(eq(test.t.a, 2), eq(test.t.b, 2)))",
           "└─Batch_Point_Get_5 2.00 root table:t, clustered index:PRIMARY(a, b, c) keep order:false, desc:false"
         ],
         "Result": [
@@ -314,8 +314,8 @@
       {
         "SQL": "select * from t where (a,b) in ((1,1),(2,2)) and c > 2;",
         "Plan": [
-          "TableReader_6 1.19 root  data:TableRangeScan_5",
-          "└─TableRangeScan_5 1.19 cop[tikv] table:t range:(1 1 2,1 1 +inf], (2 2 2,2 2 +inf], keep order:false"
+          "TableReader_6 1.00 root  data:TableRangeScan_5",
+          "└─TableRangeScan_5 1.00 cop[tikv] table:t range:(1 1 2,1 1 +inf], (2 2 2,2 2 +inf], keep order:false"
         ],
         "Result": [
           "2 2 3"
@@ -324,8 +324,8 @@
       {
         "SQL": "select * from t where ((a = 1 and b = 1) or (a = 2 and b = 2)) and c > 2;",
         "Plan": [
-          "TableReader_6 1.19 root  data:TableRangeScan_5",
-          "└─TableRangeScan_5 1.19 cop[tikv] table:t range:(1 1 2,1 1 +inf], (2 2 2,2 2 +inf], keep order:false"
+          "TableReader_6 1.00 root  data:TableRangeScan_5",
+          "└─TableRangeScan_5 1.00 cop[tikv] table:t range:(1 1 2,1 1 +inf], (2 2 2,2 2 +inf], keep order:false"
         ],
         "Result": [
           "2 2 3"
@@ -364,8 +364,8 @@
       {
         "SQL": "select * from t2 where t='aaaa';",
         "Plan": [
-          "TableReader_7 1.00 root  data:Selection_6",
-          "└─Selection_6 1.00 cop[tikv]  eq(test.t2.t, \"aaaa\")",
+          "TableReader_7 0.00 root  data:Selection_6",
+          "└─Selection_6 0.00 cop[tikv]  eq(test.t2.t, \"aaaa\")",
           "  └─TableRangeScan_5 2.00 cop[tikv] table:t2 range:[0,+inf], keep order:false"
         ],
         "Result": [
@@ -375,8 +375,8 @@
       {
         "SQL": "select * from t2 where t='aaaa' or t = 'a';",
         "Plan": [
-          "TableReader_7 1.60 root  data:Selection_6",
-          "└─Selection_6 1.60 cop[tikv]  or(eq(test.t2.t, \"aaaa\"), eq(test.t2.t, \"a\"))",
+          "TableReader_7 0.80 root  data:Selection_6",
+          "└─Selection_6 0.80 cop[tikv]  or(eq(test.t2.t, \"aaaa\"), eq(test.t2.t, \"a\"))",
           "  └─TableRangeScan_5 2.00 cop[tikv] table:t2 range:[0,+inf], keep order:false"
         ],
         "Result": [