executor: support building stats for fast analyze.

executor/analyze.go

@@ -24,6 +24,7 @@ import (
 	"sync/atomic"
 	"time"
 
+	"github.com/cznic/mathutil"
 	"github.com/pingcap/errors"
 	"github.com/pingcap/kvproto/pkg/debugpb"
 	"github.com/pingcap/parser/model"
@@ -190,6 +191,8 @@ func (e *AnalyzeExec) analyzeWorker(taskCh <-chan *analyzeTask, resultCh chan<-
 			task.job.Start()
 			resultCh <- analyzeIndexPushdown(task.idxExec)
 		case fastTask:
+			task.fastExec.job = task.job
+			task.job.Start()
 			for _, result := range analyzeFastExec(task.fastExec) {
 				resultCh <- result
 			}
@@ -507,7 +510,7 @@ func (e *AnalyzeColumnsExec) buildStats() (hists []*statistics.Histogram, cms []
 func analyzeFastExec(exec *AnalyzeFastExec) []analyzeResult {
 	hists, cms, err := exec.buildStats()
 	if err != nil {
-		return []analyzeResult{{Err: err}}
+		return []analyzeResult{{Err: err, job: exec.job}}
 	}
 	var results []analyzeResult
 	hasPKInfo := 0
@@ -522,6 +525,7 @@ func analyzeFastExec(exec *AnalyzeFastExec) []analyzeResult {
 				Cms:             []*statistics.CMSketch{cms[i]},
 				IsIndex:         1,
 				Count:           hists[i].NullCount,
+				job:             exec.job,
 			}
 			if hists[i].Len() > 0 {
 				idxResult.Count += hists[i].Buckets[hists[i].Len()-1].Count
@@ -535,6 +539,7 @@ func analyzeFastExec(exec *AnalyzeFastExec) []analyzeResult {
 		Hist:            hists[:hasPKInfo+len(exec.colsInfo)],
 		Cms:             cms[:hasPKInfo+len(exec.colsInfo)],
 		Count:           hist.NullCount,
+		job:             exec.job,
 	}
 	if hist.Len() > 0 {
 		colResult.Count += hist.Buckets[hist.Len()-1].Count
@@ -560,6 +565,7 @@ type AnalyzeFastExec struct {
 	idxsInfo        []*model.IndexInfo
 	concurrency     int
 	maxNumBuckets   uint64
+	tblInfo         *model.TableInfo
 	cache           *tikv.RegionCache
 	wg              *sync.WaitGroup
 	sampLocs        chan *tikv.KeyLocation
@@ -569,6 +575,7 @@ type AnalyzeFastExec struct {
 	scanTasks       []*tikv.KeyLocation
 	collectors      []*statistics.SampleCollector
 	randSeed        int64
+	job             *statistics.AnalyzeJob
 }
 
 func (e *AnalyzeFastExec) getSampRegionsRowCount(bo *tikv.Backoffer, needRebuild *bool, err *error, sampTasks *[]*AnalyzeFastTask) {
@@ -897,8 +904,34 @@ func (e *AnalyzeFastExec) handleSampTasks(bo *tikv.Backoffer, workID int, err *e
 }
 
 func (e *AnalyzeFastExec) buildHist(ID int64, collector *statistics.SampleCollector, tp *types.FieldType) (*statistics.Histogram, error) {
-	// TODO: build histogram and cmsketch here for one collector.
-	return nil, nil
+	// build collector properties.
+	collector.Samples = collector.Samples[:e.sampCursor]
+	err := collector.CalcTotalSize()
+	if err != nil {
+		return nil, err
+	}
+	collector.Count = int64(e.sampCursor)
+	data := make([][]byte, 0, len(collector.Samples))
+	for _, sample := range collector.Samples {
+		if sample.Value.IsNull() {
+			collector.NullCount++
+			continue
+		}
+		bytes, err := sample.Value.ToBytes()
+		if err != nil {
+			return nil, err
+		}
+		data = append(data, bytes)
+	}
+	rowCount := mathutil.MinInt64(domain.GetDomain(e.ctx).StatsHandle().GetTableStats(e.tblInfo).Count, int64(e.rowCount))
+	// build CMSketch
+	collector.CMSketch = statistics.NewCMSketchWithTopN(defaultCMSketchDepth, defaultCMSketchWidth, data, uint32(e.sampCursor), uint64(rowCount))
+	// build Histogram
+	hist, err := statistics.BuildColumnWithSamples(e.ctx, int64(e.maxNumBuckets), ID, collector, tp, rowCount)
+	if err != nil {
+		return nil, err
+	}
+	return hist, nil
 }
 
 func (e *AnalyzeFastExec) runTasks() ([]*statistics.Histogram, []*statistics.CMSketch, error) {
@@ -1009,6 +1042,7 @@ type AnalyzeTestFastExec struct {
 	IdxsInfo        []*model.IndexInfo
 	Concurrency     int
 	Collectors      []*statistics.SampleCollector
+	TblInfo         *model.TableInfo
 }
 
 // TestFastSample only test the fast sample in unit test.
@@ -1020,6 +1054,7 @@ func (e *AnalyzeTestFastExec) TestFastSample() error {
 	e.concurrency = e.Concurrency
 	e.physicalTableID = e.PhysicalTableID
 	e.wg = &sync.WaitGroup{}
+	e.tblInfo = e.TblInfo
 	_, _, err := e.buildStats()
 	e.Collectors = e.collectors
 	return err

executor/analyze_test.go

@@ -183,6 +183,7 @@ func (s *testSuite1) TestAnalyzeFastSample(c *C) {
 		IdxsInfo:        indicesInfo,
 		Concurrency:     1,
 		PhysicalTableID: tbl.(table.PhysicalTable).GetPhysicalID(),
+		TblInfo:         tblInfo,
 	}
 	err = mockExec.TestFastSample()
 	c.Assert(err, IsNil)
@@ -197,5 +198,87 @@ func (s *testSuite1) TestAnalyzeFastSample(c *C) {
 			vals[i] = append(vals[i], s)
 		}
 	}
-	c.Assert(fmt.Sprintln(vals), Equals, "[[0 34 35 57 4 24 6 25 58 9 10 11 12 30 14 52 29 17 44 54] [0 34 35 57 4 24 6 25 58 9 10 11 12 30 14 52 29 17 44 54]]\n")
+	c.Assert(fmt.Sprintln(vals), Equals, "[[0 4 6 9 10 11 12 14 17 24 25 29 30 34 35 44 52 54 57 58] [0 4 6 9 10 11 12 14 17 24 25 29 30 34 35 44 52 54 57 58]]\n")
+}
+
+func (s *testSuite1) TestFastAnalyze(c *C) {
+	cluster := mocktikv.NewCluster()
+	mocktikv.BootstrapWithSingleStore(cluster)
+	store, err := mockstore.NewMockTikvStore(
+		mockstore.WithCluster(cluster),
+	)
+	c.Assert(err, IsNil)
+	var dom *domain.Domain
+	dom, err = session.BootstrapSession(store)
+	c.Assert(err, IsNil)
+	tk := testkit.NewTestKit(c, store)
+	executor.MaxSampleSize = 1000
+	executor.RandSeed = 123
+
+	tk.MustExec("use test")
+	tk.MustExec("drop table if exists t")
+	tk.MustExec("create table t(a int primary key, b int, index index_b(b))")
+	tk.MustExec("set @@session.tidb_enable_fast_analyze=1")
+	tk.MustExec("set @@session.tidb_build_stats_concurrency=1")
+	for i := 0; i < 3000; i++ {
+		tk.MustExec(fmt.Sprintf("insert into t values (%d, %d)", i, i))
+	}
+	tblInfo, err := dom.InfoSchema().TableByName(model.NewCIStr("test"), model.NewCIStr("t"))
+	c.Assert(err, IsNil)
+	tid := tblInfo.Meta().ID
+
+	// construct 5 regions split by {600, 1200, 1800, 2400}
+	splitKeys := generateTableSplitKeyForInt(tid, []int{600, 1200, 1800, 2400})
+	manipulateCluster(cluster, splitKeys)
+
+	tk.MustExec("analyze table t with 5 buckets")
+
+	is := executor.GetInfoSchema(tk.Se.(sessionctx.Context))
+	table, err := is.TableByName(model.NewCIStr("test"), model.NewCIStr("t"))
+	c.Assert(err, IsNil)
+	tableInfo := table.Meta()
+	tbl := dom.StatsHandle().GetTableStats(tableInfo)
+	sTbl := fmt.Sprintln(tbl)
+	matched := false
+	if sTbl == "Table:37 Count:3000\n"+
+		"column:1 ndv:3000 totColSize:3639\n"+
+		"num: 603 lower_bound: 6 upper_bound: 612 repeats: 1\n"+
+		"num: 603 lower_bound: 621 upper_bound: 1205 repeats: 1\n"+
+		"num: 603 lower_bound: 1207 upper_bound: 1830 repeats: 1\n"+
+		"num: 603 lower_bound: 1831 upper_bound: 2387 repeats: 1\n"+
+		"num: 588 lower_bound: 2390 upper_bound: 2997 repeats: 1\n"+
+		"column:2 ndv:3000 totColSize:3639\n"+
+		"num: 603 lower_bound: 6 upper_bound: 612 repeats: 1\n"+
+		"num: 603 lower_bound: 621 upper_bound: 1205 repeats: 1\n"+
+		"num: 603 lower_bound: 1207 upper_bound: 1830 repeats: 1\n"+
+		"num: 603 lower_bound: 1831 upper_bound: 2387 repeats: 1\n"+
+		"num: 588 lower_bound: 2390 upper_bound: 2997 repeats: 1\n"+
+		"index:1 ndv:3000\n"+
+		"num: 603 lower_bound: 6 upper_bound: 612 repeats: 1\n"+
+		"num: 603 lower_bound: 621 upper_bound: 1205 repeats: 1\n"+
+		"num: 603 lower_bound: 1207 upper_bound: 1830 repeats: 1\n"+
+		"num: 603 lower_bound: 1831 upper_bound: 2387 repeats: 1\n"+
+		"num: 588 lower_bound: 2390 upper_bound: 2997 repeats: 1\n" ||
+		sTbl == "Table:37 Count:3000\n"+
+			"column:2 ndv:3000 totColSize:3639\n"+
+			"num: 603 lower_bound: 6 upper_bound: 612 repeats: 1\n"+
+			"num: 603 lower_bound: 621 upper_bound: 1205 repeats: 1\n"+
+			"num: 603 lower_bound: 1207 upper_bound: 1830 repeats: 1\n"+
+			"num: 603 lower_bound: 1831 upper_bound: 2387 repeats: 1\n"+
+			"num: 588 lower_bound: 2390 upper_bound: 2997 repeats: 1\n"+
+			"column:1 ndv:3000 totColSize:3639\n"+
+			"num: 603 lower_bound: 6 upper_bound: 612 repeats: 1\n"+
+			"num: 603 lower_bound: 621 upper_bound: 1205 repeats: 1\n"+
+			"num: 603 lower_bound: 1207 upper_bound: 1830 repeats: 1\n"+
+			"num: 603 lower_bound: 1831 upper_bound: 2387 repeats: 1\n"+
+			"num: 588 lower_bound: 2390 upper_bound: 2997 repeats: 1\n"+
+			"index:1 ndv:3000\n"+
+			"num: 603 lower_bound: 6 upper_bound: 612 repeats: 1\n"+
+			"num: 603 lower_bound: 621 upper_bound: 1205 repeats: 1\n"+
+			"num: 603 lower_bound: 1207 upper_bound: 1830 repeats: 1\n"+
+			"num: 603 lower_bound: 1831 upper_bound: 2387 repeats: 1\n"+
+			"num: 588 lower_bound: 2390 upper_bound: 2997 repeats: 1\n" {
+		matched = true
+	}
+	c.Assert(matched, Equals, true)
 }

executor/builder.go

@@ -1435,9 +1435,11 @@ func (b *executorBuilder) buildAnalyzeFastColumn(e *AnalyzeExec, task plannercor
 				colsInfo:        task.ColsInfo,
 				pkInfo:          task.PKInfo,
 				maxNumBuckets:   maxNumBuckets,
+				tblInfo:         task.TblInfo,
 				concurrency:     concurrency,
 				wg:              &sync.WaitGroup{},
 			},
+			job: &statistics.AnalyzeJob{DBName: task.DBName, TableName: task.TableName, PartitionName: task.PartitionName, JobInfo: "fast analyze columns"},
 		})
 	}
 }
@@ -1464,9 +1466,11 @@ func (b *executorBuilder) buildAnalyzeFastIndex(e *AnalyzeExec, task plannercore
 				physicalTableID: task.PhysicalTableID,
 				idxsInfo:        []*model.IndexInfo{task.IndexInfo},
 				maxNumBuckets:   maxNumBuckets,
+				tblInfo:         task.TblInfo,
 				concurrency:     concurrency,
 				wg:              &sync.WaitGroup{},
 			},
+			job: &statistics.AnalyzeJob{DBName: task.DBName, TableName: task.TableName, PartitionName: "fast analyze index " + task.IndexInfo.Name.O},
 		})
 	}
 }

planner/core/common_plans.go

@@ -427,20 +427,20 @@ type analyzeInfo struct {
 	PartitionName string
 	// PhysicalTableID is the id for a partition or a table.
 	PhysicalTableID int64
-	PKInfo          *model.ColumnInfo
-	ColsInfo        []*model.ColumnInfo
 }
 
 // AnalyzeColumnsTask is used for analyze columns.
 type AnalyzeColumnsTask struct {
 	PKInfo   *model.ColumnInfo
 	ColsInfo []*model.ColumnInfo
+	TblInfo  *model.TableInfo
 	analyzeInfo
 }
 
 // AnalyzeIndexTask is used for analyze index.
 type AnalyzeIndexTask struct {
 	IndexInfo *model.IndexInfo
+	TblInfo   *model.TableInfo
 	analyzeInfo
 }
 

planner/core/planbuilder.go

@@ -812,6 +812,7 @@ func (b *PlanBuilder) buildAnalyzeTable(as *ast.AnalyzeTableStmt) (Plan, error)
 				p.IdxTasks = append(p.IdxTasks, AnalyzeIndexTask{
 					IndexInfo:   idx,
 					analyzeInfo: info,
+					TblInfo:     tbl.TableInfo,
 				})
 			}
 		}
@@ -822,6 +823,7 @@ func (b *PlanBuilder) buildAnalyzeTable(as *ast.AnalyzeTableStmt) (Plan, error)
 					PKInfo:      pkInfo,
 					ColsInfo:    colInfo,
 					analyzeInfo: info,
+					TblInfo:     tbl.TableInfo,
 				})
 			}
 		}
@@ -843,7 +845,7 @@ func (b *PlanBuilder) buildAnalyzeIndex(as *ast.AnalyzeTableStmt) (Plan, error)
 		}
 		for i, id := range physicalIDs {
 			info := analyzeInfo{DBName: as.TableNames[0].Schema.O, TableName: as.TableNames[0].Name.O, PartitionName: names[i], PhysicalTableID: id}
-			p.IdxTasks = append(p.IdxTasks, AnalyzeIndexTask{IndexInfo: idx, analyzeInfo: info})
+			p.IdxTasks = append(p.IdxTasks, AnalyzeIndexTask{IndexInfo: idx, analyzeInfo: info, TblInfo: tblInfo})
 		}
 	}
 	return p, nil
@@ -860,7 +862,7 @@ func (b *PlanBuilder) buildAnalyzeAllIndex(as *ast.AnalyzeTableStmt) (Plan, erro
 		if idx.State == model.StatePublic {
 			for i, id := range physicalIDs {
 				info := analyzeInfo{DBName: as.TableNames[0].Schema.O, TableName: as.TableNames[0].Name.O, PartitionName: names[i], PhysicalTableID: id}
-				p.IdxTasks = append(p.IdxTasks, AnalyzeIndexTask{IndexInfo: idx, analyzeInfo: info})
+				p.IdxTasks = append(p.IdxTasks, AnalyzeIndexTask{IndexInfo: idx, analyzeInfo: info, TblInfo: tblInfo})
 			}
 		}
 	}

statistics/builder.go

@@ -93,23 +93,28 @@ func (b *SortedBuilder) Iterate(data types.Datum) error {
 	return nil
 }
 
-// BuildColumn builds histogram from samples for column.
-func BuildColumn(ctx sessionctx.Context, numBuckets, id int64, collector *SampleCollector, tp *types.FieldType) (*Histogram, error) {
-	count := collector.Count
-	ndv := collector.FMSketch.NDV()
+// buildColumnHist build a histogram for a table.
+// numBuckets: number of buckets for the histogram.
+// id: the id of the table.
+// collector: the collector of samples.
+// tp: the FieldType for the column.
+// count: represents the row count for the column.
+// ndv: represents the number of distinct values for the column.
+// nullCount: represents the number of null values for the column.
+func buildColumnHist(ctx sessionctx.Context, numBuckets, id int64, collector *SampleCollector, tp *types.FieldType, count int64, ndv int64, nullCount int64) (*Histogram, error) {
 	if ndv > count {
 		ndv = count
 	}
 	if count == 0 || len(collector.Samples) == 0 {
-		return NewHistogram(id, ndv, collector.NullCount, 0, tp, 0, collector.TotalSize), nil
+		return NewHistogram(id, ndv, nullCount, 0, tp, 0, collector.TotalSize), nil
 	}
 	sc := ctx.GetSessionVars().StmtCtx
 	samples := collector.Samples
 	err := SortSampleItems(sc, samples)
 	if err != nil {
 		return nil, err
 	}
-	hg := NewHistogram(id, ndv, collector.NullCount, 0, tp, int(numBuckets), collector.TotalSize)
+	hg := NewHistogram(id, ndv, nullCount, 0, tp, int(numBuckets), collector.TotalSize)
 
 	sampleNum := int64(len(samples))
 	// As we use samples to build the histogram, the bucket number and repeat should multiply a factor.
@@ -174,3 +179,24 @@ func BuildColumn(ctx sessionctx.Context, numBuckets, id int64, collector *Sample
 	hg.Correlation = (itemsCount*corrXYSum - corrXSum*corrXSum) / (itemsCount*corrX2Sum - corrXSum*corrXSum)
 	return hg, nil
 }
+
+// BuildColumn builds histogram from samples for column.
+func BuildColumn(ctx sessionctx.Context, numBuckets, id int64, collector *SampleCollector, tp *types.FieldType) (*Histogram, error) {
+	return buildColumnHist(ctx, numBuckets, id, collector, tp, collector.Count, collector.FMSketch.NDV(), collector.NullCount)
+}
+
+// BuildColumnWithSamples builds histogram from samples for column.
+// It was used in that collector.Count is not the entire count but the sample count.
+func BuildColumnWithSamples(ctx sessionctx.Context, numBuckets, id int64, collector *SampleCollector, tp *types.FieldType, count int64) (*Histogram, error) {
+	samplesBytes := make([][]byte, 0, len(collector.Samples))
+	for _, sample := range collector.Samples {
+		bytes, err := sample.Value.ToBytes()
+		if err != nil {
+			return nil, err
+		}
+		samplesBytes = append(samplesBytes, bytes)
+	}
+	ndv, scaleRatio := calculateEstimateNDV(newTopNHelper(samplesBytes, 0), uint64(count))
+	nullCount := collector.NullCount * int64(scaleRatio)
+	return buildColumnHist(ctx, numBuckets, id, collector, tp, count, int64(ndv), nullCount)
+}

statistics/sample.go

@@ -17,6 +17,7 @@ import (
 	"context"
 	"math/rand"
 	"sort"
+	"unicode/utf8"
 
 	"github.com/pingcap/errors"
 	"github.com/pingcap/parser/ast"
@@ -173,6 +174,19 @@ func (c *SampleCollector) collect(sc *stmtctx.StatementContext, d types.Datum) e
 	return nil
 }
 
+// CalcTotalSize is to calculate total size based on samples.
+func (c *SampleCollector) CalcTotalSize() error {
+	c.TotalSize = 0
+	for _, item := range c.Samples {
+		str, err := item.Value.ToString()
+		if err != nil {
+			return err
+		}
+		c.TotalSize += int64(utf8.RuneCountInString(str))
+	}
+	return nil
+}
+
 // SampleBuilder is used to build samples for columns.
 // Also, if primary key is handle, it will directly build histogram for it.
 type SampleBuilder struct {