statistics: add bucket ndv for index histogram

cmd/explaintest/run-tests.sh

@@ -208,7 +208,7 @@ if [ "${TIDB_TEST_STORE_NAME}" = "tikv" ]; then
     $tidb_server -P "$port" -status "$status" -config config.toml -store tikv -path "${TIKV_PATH}" > $explain_test_log 2>&1 &
     SERVER_PID=$!
 else
-    $tidb_server -P "$port" -status "$status" -config config.toml -store mocktikv -path "" > $explain_test_log 2>&1 &
+    $tidb_server -P "$port" -status "$status" -config config.toml -store unistore -path "" > $explain_test_log 2>&1 &
     SERVER_PID=$!
 fi
 echo "tidb-server(PID: $SERVER_PID) started"

distsql/select_result.go

@@ -147,9 +147,8 @@ func (r *selectResult) fetchResp(ctx context.Context) error {
 			sc.AppendWarning(dbterror.ClassTiKV.Synthesize(terror.ErrCode(warning.Code), warning.Msg))
 		}
 		if r.feedback != nil {
-			r.feedback.Update(resultSubset.GetStartKey(), r.selectResp.OutputCounts)
+			r.feedback.Update(resultSubset.GetStartKey(), r.selectResp.OutputCounts, r.selectResp.Ndvs)
 		}
-
 		r.partialCount++
 
 		hasStats, ok := resultSubset.(CopRuntimeStats)

distsql/stream.go

@@ -105,7 +105,7 @@ func (r *streamResult) readDataFromResponse(ctx context.Context, resp kv.Respons
 	if err != nil {
 		return false, errors.Trace(err)
 	}
-	r.feedback.Update(resultSubset.GetStartKey(), stream.OutputCounts)
+	r.feedback.Update(resultSubset.GetStartKey(), stream.OutputCounts, stream.Ndvs)
 	r.partialCount++
 
 	hasStats, ok := resultSubset.(CopRuntimeStats)

executor/analyze.go

@@ -335,6 +335,10 @@ func (e *AnalyzeIndexExec) buildStatsFromResult(result distsql.SelectResult, nee
 		cms = statistics.NewCMSketch(int32(e.opts[ast.AnalyzeOptCMSketchDepth]), int32(e.opts[ast.AnalyzeOptCMSketchWidth]))
 		topn = statistics.NewTopN(int(e.opts[ast.AnalyzeOptNumTopN]))
 	}
+	statsVer := statistics.Version1
+	if e.analyzePB.IdxReq.Version != nil {
+		statsVer = int(*e.analyzePB.IdxReq.Version)
+	}
 	for {
 		data, err := result.NextRaw(context.TODO())
 		if err != nil {
@@ -350,7 +354,7 @@ func (e *AnalyzeIndexExec) buildStatsFromResult(result distsql.SelectResult, nee
 		}
 		respHist := statistics.HistogramFromProto(resp.Hist)
 		e.job.Update(int64(respHist.TotalRowCount()))
-		hist, err = statistics.MergeHistograms(e.ctx.GetSessionVars().StmtCtx, hist, respHist, int(e.opts[ast.AnalyzeOptNumBuckets]))
+		hist, err = statistics.MergeHistograms(e.ctx.GetSessionVars().StmtCtx, hist, respHist, int(e.opts[ast.AnalyzeOptNumBuckets]), statsVer)
 		if err != nil {
 			return nil, nil, nil, err
 		}
@@ -556,7 +560,7 @@ func (e *AnalyzeColumnsExec) buildStats(ranges []*ranger.Range, needExtStats boo
 		if hasPkHist(e.handleCols) {
 			respHist := statistics.HistogramFromProto(resp.PkHist)
 			rowCount = int64(respHist.TotalRowCount())
-			pkHist, err = statistics.MergeHistograms(sc, pkHist, respHist, int(e.opts[ast.AnalyzeOptNumBuckets]))
+			pkHist, err = statistics.MergeHistograms(sc, pkHist, respHist, int(e.opts[ast.AnalyzeOptNumBuckets]), statistics.Version1)
 			if err != nil {
 				return nil, nil, nil, nil, err
 			}
@@ -1244,7 +1248,7 @@ func analyzeIndexIncremental(idxExec *analyzeIndexIncrementalExec) analyzeResult
 	if err != nil {
 		return analyzeResult{Err: err, job: idxExec.job}
 	}
-	hist, err = statistics.MergeHistograms(idxExec.ctx.GetSessionVars().StmtCtx, idxExec.oldHist, hist, int(idxExec.opts[ast.AnalyzeOptNumBuckets]))
+	hist, err = statistics.MergeHistograms(idxExec.ctx.GetSessionVars().StmtCtx, idxExec.oldHist, hist, int(idxExec.opts[ast.AnalyzeOptNumBuckets]), statistics.Version1)
 	if err != nil {
 		return analyzeResult{Err: err, job: idxExec.job}
 	}
@@ -1295,7 +1299,7 @@ func analyzePKIncremental(colExec *analyzePKIncrementalExec) analyzeResult {
 		return analyzeResult{Err: err, job: colExec.job}
 	}
 	hist := hists[0]
-	hist, err = statistics.MergeHistograms(colExec.ctx.GetSessionVars().StmtCtx, colExec.oldHist, hist, int(colExec.opts[ast.AnalyzeOptNumBuckets]))
+	hist, err = statistics.MergeHistograms(colExec.ctx.GetSessionVars().StmtCtx, colExec.oldHist, hist, int(colExec.opts[ast.AnalyzeOptNumBuckets]), statistics.Version1)
 	if err != nil {
 		return analyzeResult{Err: err, job: colExec.job}
 	}

executor/analyze_test.go

@@ -461,8 +461,8 @@ func (s *testFastAnalyze) TestFastAnalyze(c *C) {
 	tk.MustExec("insert into t2 values (0), (18446744073709551615)")
 	tk.MustExec("analyze table t2")
 	tk.MustQuery("show stats_buckets where table_name = 't2'").Check(testkit.Rows(
-		"test t2  a 0 0 1 1 0 0",
-		"test t2  a 0 1 2 1 18446744073709551615 18446744073709551615"))
+		"test t2  a 0 0 1 1 0 0 0",
+		"test t2  a 0 1 2 1 18446744073709551615 18446744073709551615 0"))
 
 	tk.MustExec(`set @@tidb_partition_prune_mode='` + string(variable.StaticOnly) + `'`)
 	tk.MustExec(`create table t3 (id int, v int, primary key(id), index k(v)) partition by hash (id) partitions 4`)
@@ -531,6 +531,7 @@ func (s *testSuite1) TestAnalyzeIncremental(c *C) {
 }
 
 func (s *testSuite1) TestAnalyzeIncrementalStreaming(c *C) {
+	c.Skip("unistore hasn't support streaming yet.")
 	tk := testkit.NewTestKit(c, s.store)
 	tk.MustExec("use test")
 	tk.Se.GetSessionVars().EnableStreaming = true
@@ -545,13 +546,13 @@ func (s *testSuite1) testAnalyzeIncremental(tk *testkit.TestKit, c *C) {
 	tk.MustQuery("show stats_buckets").Check(testkit.Rows())
 	tk.MustExec("insert into t values (1,1)")
 	tk.MustExec("analyze incremental table t index")
-	tk.MustQuery("show stats_buckets").Check(testkit.Rows("test t  a 0 0 1 1 1 1", "test t  idx 1 0 1 1 1 1"))
+	tk.MustQuery("show stats_buckets").Check(testkit.Rows("test t  a 0 0 1 1 1 1 0", "test t  idx 1 0 1 1 1 1 0"))
 	tk.MustExec("insert into t values (2,2)")
 	tk.MustExec("analyze incremental table t index")
-	tk.MustQuery("show stats_buckets").Check(testkit.Rows("test t  a 0 0 1 1 1 1", "test t  a 0 1 2 1 2 2", "test t  idx 1 0 1 1 1 1", "test t  idx 1 1 2 1 2 2"))
+	tk.MustQuery("show stats_buckets").Check(testkit.Rows("test t  a 0 0 1 1 1 1 0", "test t  a 0 1 2 1 2 2 0", "test t  idx 1 0 1 1 1 1 0", "test t  idx 1 1 2 1 2 2 0"))
 	tk.MustExec("analyze incremental table t index")
 	// Result should not change.
-	tk.MustQuery("show stats_buckets").Check(testkit.Rows("test t  a 0 0 1 1 1 1", "test t  a 0 1 2 1 2 2", "test t  idx 1 0 1 1 1 1", "test t  idx 1 1 2 1 2 2"))
+	tk.MustQuery("show stats_buckets").Check(testkit.Rows("test t  a 0 0 1 1 1 1 0", "test t  a 0 1 2 1 2 2 0", "test t  idx 1 0 1 1 1 1 0", "test t  idx 1 1 2 1 2 2 0"))
 
 	// Test analyze incremental with feedback.
 	tk.MustExec("insert into t values (3,3)")
@@ -574,7 +575,7 @@ func (s *testSuite1) testAnalyzeIncremental(tk *testkit.TestKit, c *C) {
 	c.Assert(h.DumpStatsFeedbackToKV(), IsNil)
 	c.Assert(h.HandleUpdateStats(is), IsNil)
 	c.Assert(h.Update(is), IsNil)
-	tk.MustQuery("show stats_buckets").Check(testkit.Rows("test t  a 0 0 1 1 1 1", "test t  a 0 1 3 0 2 2147483647", "test t  idx 1 0 1 1 1 1", "test t  idx 1 1 2 1 2 2"))
+	tk.MustQuery("show stats_buckets").Check(testkit.Rows("test t  a 0 0 1 1 1 1 0", "test t  a 0 1 3 0 2 2147483647 0", "test t  idx 1 0 1 1 1 1 0", "test t  idx 1 1 2 1 2 2 0"))
 	tblStats := h.GetTableStats(tblInfo)
 	val, err := codec.EncodeKey(tk.Se.GetSessionVars().StmtCtx, nil, types.NewIntDatum(3))
 	c.Assert(err, IsNil)
@@ -583,8 +584,8 @@ func (s *testSuite1) testAnalyzeIncremental(tk *testkit.TestKit, c *C) {
 	c.Assert(statistics.IsAnalyzed(tblStats.Columns[tblInfo.Columns[0].ID].Flag), IsFalse)
 
 	tk.MustExec("analyze incremental table t index")
-	tk.MustQuery("show stats_buckets").Check(testkit.Rows("test t  a 0 0 1 1 1 1", "test t  a 0 1 2 1 2 2", "test t  a 0 2 3 1 3 3",
-		"test t  idx 1 0 1 1 1 1", "test t  idx 1 1 2 1 2 2", "test t  idx 1 2 3 1 3 3"))
+	tk.MustQuery("show stats_buckets").Check(testkit.Rows("test t  a 0 0 1 1 1 1 0", "test t  a 0 1 2 1 2 2 0", "test t  a 0 2 3 1 3 3 0",
+		"test t  idx 1 0 1 1 1 1 0", "test t  idx 1 1 2 1 2 2 0", "test t  idx 1 2 3 1 3 3 0"))
 	tblStats = h.GetTableStats(tblInfo)
 	c.Assert(tblStats.Indices[tblInfo.Indices[0].ID].QueryBytes(val), Equals, uint64(1))
 }
@@ -769,36 +770,36 @@ func (s *testSuite1) TestNormalAnalyzeOnCommonHandle(c *C) {
 	tk.MustExec("analyze table t1, t2, t3")
 
 	tk.MustQuery(`show stats_buckets where table_name in ("t1", "t2", "t3")`).Sort().Check(testkit.Rows(
-		"test t1  a 0 0 1 1 1 1",
-		"test t1  a 0 1 2 1 2 2",
-		"test t1  a 0 2 3 1 3 3",
-		"test t1  b 0 0 1 1 1 1",
-		"test t1  b 0 1 2 1 2 2",
-		"test t1  b 0 2 3 1 3 3",
-		"test t2  PRIMARY 1 0 1 1 111 111",
-		"test t2  PRIMARY 1 1 2 1 222 222",
-		"test t2  PRIMARY 1 2 3 1 333 333",
-		"test t2  a 0 0 1 1 111 111",
-		"test t2  a 0 1 2 1 222 222",
-		"test t2  a 0 2 3 1 333 333",
-		"test t2  b 0 0 1 1 1 1",
-		"test t2  b 0 1 2 1 2 2",
-		"test t2  b 0 2 3 1 3 3",
-		"test t3  PRIMARY 1 0 1 1 (1, 1) (1, 1)",
-		"test t3  PRIMARY 1 1 2 1 (2, 2) (2, 2)",
-		"test t3  PRIMARY 1 2 3 1 (3, 3) (3, 3)",
-		"test t3  a 0 0 1 1 1 1",
-		"test t3  a 0 1 2 1 2 2",
-		"test t3  a 0 2 3 1 3 3",
-		"test t3  b 0 0 1 1 1 1",
-		"test t3  b 0 1 2 1 2 2",
-		"test t3  b 0 2 3 1 3 3",
-		"test t3  c 0 0 1 1 1 1",
-		"test t3  c 0 1 2 1 2 2",
-		"test t3  c 0 2 3 1 3 3",
-		"test t3  c 1 0 1 1 1 1",
-		"test t3  c 1 1 2 1 2 2",
-		"test t3  c 1 2 3 1 3 3"))
+		"test t1  a 0 0 1 1 1 1 0",
+		"test t1  a 0 1 2 1 2 2 0",
+		"test t1  a 0 2 3 1 3 3 0",
+		"test t1  b 0 0 1 1 1 1 0",
+		"test t1  b 0 1 2 1 2 2 0",
+		"test t1  b 0 2 3 1 3 3 0",
+		"test t2  PRIMARY 1 0 1 1 111 111 0",
+		"test t2  PRIMARY 1 1 2 1 222 222 0",
+		"test t2  PRIMARY 1 2 3 1 333 333 0",
+		"test t2  a 0 0 1 1 111 111 0",
+		"test t2  a 0 1 2 1 222 222 0",
+		"test t2  a 0 2 3 1 333 333 0",
+		"test t2  b 0 0 1 1 1 1 0",
+		"test t2  b 0 1 2 1 2 2 0",
+		"test t2  b 0 2 3 1 3 3 0",
+		"test t3  PRIMARY 1 0 1 1 (1, 1) (1, 1) 0",
+		"test t3  PRIMARY 1 1 2 1 (2, 2) (2, 2) 0",
+		"test t3  PRIMARY 1 2 3 1 (3, 3) (3, 3) 0",
+		"test t3  a 0 0 1 1 1 1 0",
+		"test t3  a 0 1 2 1 2 2 0",
+		"test t3  a 0 2 3 1 3 3 0",
+		"test t3  b 0 0 1 1 1 1 0",
+		"test t3  b 0 1 2 1 2 2 0",
+		"test t3  b 0 2 3 1 3 3 0",
+		"test t3  c 0 0 1 1 1 1 0",
+		"test t3  c 0 1 2 1 2 2 0",
+		"test t3  c 0 2 3 1 3 3 0",
+		"test t3  c 1 0 1 1 1 1 0",
+		"test t3  c 1 1 2 1 2 2 0",
+		"test t3  c 1 2 3 1 3 3 0"))
 }
 
 func (s *testSuite1) TestDefaultValForAnalyze(c *C) {
@@ -846,15 +847,15 @@ func (s *testSerialSuite2) TestIssue20874(c *C) {
 	tk.MustExec("insert into t values ('#', 'C'), ('$', 'c'), ('a', 'a')")
 	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",
-		"test t  a 0 1 2 1 \x0e\x0f \x0e\x0f",
-		"test t  a 0 2 3 1 \x0e3 \x0e3",
-		"test t  b 0 0 1 1 \x00A \x00A",
-		"test t  b 0 1 3 2 \x00C \x00C",
-		"test t  idxa 1 0 1 1 \x02\xd2 \x02\xd2",
-		"test t  idxa 1 1 2 1 \x0e\x0f \x0e\x0f",
-		"test t  idxa 1 2 3 1 \x0e3 \x0e3",
-		"test t  idxb 1 0 1 1 \x00A \x00A",
-		"test t  idxb 1 1 3 2 \x00C \x00C",
+		"test t  a 0 0 1 1 \x02\xd2 \x02\xd2 0",
+		"test t  a 0 1 2 1 \x0e\x0f \x0e\x0f 0",
+		"test t  a 0 2 3 1 \x0e3 \x0e3 0",
+		"test t  b 0 0 1 1 \x00A \x00A 0",
+		"test t  b 0 1 3 2 \x00C \x00C 0",
+		"test t  idxa 1 0 1 1 \x02\xd2 \x02\xd2 0",
+		"test t  idxa 1 1 2 1 \x0e\x0f \x0e\x0f 0",
+		"test t  idxa 1 2 3 1 \x0e3 \x0e3 0",
+		"test t  idxb 1 0 1 1 \x00A \x00A 0",
+		"test t  idxb 1 1 3 2 \x00C \x00C 0",
 	))
 }

executor/show_stats.go

@@ -259,6 +259,7 @@ func (e *ShowExec) bucketsToRows(dbName, tblName, partitionName, colName string,
 			hist.Buckets[i].Repeat,
 			lowerBoundStr,
 			upperBoundStr,
+			hist.Buckets[i].NDV,
 		})
 	}
 	return nil

executor/show_stats_test.go

@@ -80,36 +80,36 @@ func (s *testShowStatsSuite) TestShowStatsBuckets(c *C) {
 	tk.MustExec("insert into t values (1,1)")
 	tk.MustExec("analyze table t")
 	result := tk.MustQuery("show stats_buckets").Sort()
-	result.Check(testkit.Rows("test t  a 0 0 1 1 1 1", "test t  b 0 0 1 1 1 1", "test t  idx 1 0 1 1 (1, 1) (1, 1)"))
+	result.Check(testkit.Rows("test t  a 0 0 1 1 1 1 0", "test t  b 0 0 1 1 1 1 0", "test t  idx 1 0 1 1 (1, 1) (1, 1) 0"))
 	result = tk.MustQuery("show stats_buckets where column_name = 'idx'")
-	result.Check(testkit.Rows("test t  idx 1 0 1 1 (1, 1) (1, 1)"))
+	result.Check(testkit.Rows("test t  idx 1 0 1 1 (1, 1) (1, 1) 0"))
 
 	tk.MustExec("drop table t")
 	tk.MustExec("create table t (`a` datetime, `b` int, key `idx`(`a`, `b`))")
 	tk.MustExec("insert into t values (\"2020-01-01\", 1)")
 	tk.MustExec("analyze table t")
 	result = tk.MustQuery("show stats_buckets").Sort()
-	result.Check(testkit.Rows("test t  a 0 0 1 1 2020-01-01 00:00:00 2020-01-01 00:00:00", "test t  b 0 0 1 1 1 1", "test t  idx 1 0 1 1 (2020-01-01 00:00:00, 1) (2020-01-01 00:00:00, 1)"))
+	result.Check(testkit.Rows("test t  a 0 0 1 1 2020-01-01 00:00:00 2020-01-01 00:00:00 0", "test t  b 0 0 1 1 1 1 0", "test t  idx 1 0 1 1 (2020-01-01 00:00:00, 1) (2020-01-01 00:00:00, 1) 0"))
 	result = tk.MustQuery("show stats_buckets where column_name = 'idx'")
-	result.Check(testkit.Rows("test t  idx 1 0 1 1 (2020-01-01 00:00:00, 1) (2020-01-01 00:00:00, 1)"))
+	result.Check(testkit.Rows("test t  idx 1 0 1 1 (2020-01-01 00:00:00, 1) (2020-01-01 00:00:00, 1) 0"))
 
 	tk.MustExec("drop table t")
 	tk.MustExec("create table t (`a` date, `b` int, key `idx`(`a`, `b`))")
 	tk.MustExec("insert into t values (\"2020-01-01\", 1)")
 	tk.MustExec("analyze table t")
 	result = tk.MustQuery("show stats_buckets").Sort()
-	result.Check(testkit.Rows("test t  a 0 0 1 1 2020-01-01 2020-01-01", "test t  b 0 0 1 1 1 1", "test t  idx 1 0 1 1 (2020-01-01, 1) (2020-01-01, 1)"))
+	result.Check(testkit.Rows("test t  a 0 0 1 1 2020-01-01 2020-01-01 0", "test t  b 0 0 1 1 1 1 0", "test t  idx 1 0 1 1 (2020-01-01, 1) (2020-01-01, 1) 0"))
 	result = tk.MustQuery("show stats_buckets where column_name = 'idx'")
-	result.Check(testkit.Rows("test t  idx 1 0 1 1 (2020-01-01, 1) (2020-01-01, 1)"))
+	result.Check(testkit.Rows("test t  idx 1 0 1 1 (2020-01-01, 1) (2020-01-01, 1) 0"))
 
 	tk.MustExec("drop table t")
 	tk.MustExec("create table t (`a` timestamp, `b` int, key `idx`(`a`, `b`))")
 	tk.MustExec("insert into t values (\"2020-01-01\", 1)")
 	tk.MustExec("analyze table t")
 	result = tk.MustQuery("show stats_buckets").Sort()
-	result.Check(testkit.Rows("test t  a 0 0 1 1 2020-01-01 00:00:00 2020-01-01 00:00:00", "test t  b 0 0 1 1 1 1", "test t  idx 1 0 1 1 (2020-01-01 00:00:00, 1) (2020-01-01 00:00:00, 1)"))
+	result.Check(testkit.Rows("test t  a 0 0 1 1 2020-01-01 00:00:00 2020-01-01 00:00:00 0", "test t  b 0 0 1 1 1 1 0", "test t  idx 1 0 1 1 (2020-01-01 00:00:00, 1) (2020-01-01 00:00:00, 1) 0"))
 	result = tk.MustQuery("show stats_buckets where column_name = 'idx'")
-	result.Check(testkit.Rows("test t  idx 1 0 1 1 (2020-01-01 00:00:00, 1) (2020-01-01 00:00:00, 1)"))
+	result.Check(testkit.Rows("test t  idx 1 0 1 1 (2020-01-01 00:00:00, 1) (2020-01-01 00:00:00, 1) 0"))
 }
 
 func (s *testShowStatsSuite) TestShowStatsHasNullValue(c *C) {
@@ -124,14 +124,14 @@ func (s *testShowStatsSuite) TestShowStatsHasNullValue(c *C) {
 	tk.MustExec("insert into t values(1)")
 	tk.MustExec("analyze table t")
 	tk.MustQuery("show stats_buckets").Sort().Check(testkit.Rows(
-		"test t  a 0 0 1 1 1 1",
-		"test t  idx 1 0 1 1 1 1",
+		"test t  a 0 0 1 1 1 1 0",
+		"test t  idx 1 0 1 1 1 1 0",
 	))
 	tk.MustExec("drop table t")
 	tk.MustExec("create table t (a int, b int, index idx(a, b))")
 	tk.MustExec("insert into t values(NULL, NULL)")
 	tk.MustExec("analyze table t")
-	tk.MustQuery("show stats_buckets").Check(testkit.Rows("test t  idx 1 0 1 1 (NULL, NULL) (NULL, NULL)"))
+	tk.MustQuery("show stats_buckets").Check(testkit.Rows("test t  idx 1 0 1 1 (NULL, NULL) (NULL, NULL) 0"))
 
 	tk.MustExec("drop table t")
 	tk.MustExec("create table t(a int, b int, c int, index idx_b(b), index idx_c_a(c, a))")
@@ -201,7 +201,7 @@ func (s *testShowStatsSuite) TestShowPartitionStats(c *C) {
 		c.Assert(result.Rows()[2][3], Equals, "idx")
 
 		result = tk.MustQuery("show stats_buckets").Sort()
-		result.Check(testkit.Rows("test t p0 a 0 0 1 1 1 1", "test t p0 b 0 0 1 1 1 1", "test t p0 idx 1 0 1 1 1 1"))
+		result.Check(testkit.Rows("test t p0 a 0 0 1 1 1 1 0", "test t p0 b 0 0 1 1 1 1 0", "test t p0 idx 1 0 1 1 1 1 0"))
 
 		result = tk.MustQuery("show stats_healthy")
 		result.Check(testkit.Rows("test t p0 100"))

go.mod

@@ -51,7 +51,7 @@ require (
 	github.com/pingcap/sysutil v0.0.0-20201130064824-f0c8aa6a6966
 	github.com/pingcap/tidb-lightning v4.0.9-0.20201106041742-a1ac97827a27+incompatible
 	github.com/pingcap/tidb-tools v4.0.9-0.20201127090955-2707c97b3853+incompatible
-	github.com/pingcap/tipb v0.0.0-20201209065231-aa39b1b86217
+	github.com/pingcap/tipb v0.0.0-20201215091753-bd0cb2b314a4
 	github.com/prometheus/client_golang v1.5.1
 	github.com/prometheus/client_model v0.2.0
 	github.com/prometheus/common v0.9.1

go.sum

@@ -711,8 +711,8 @@ github.com/pingcap/tidb-tools v4.0.9-0.20201127090955-2707c97b3853+incompatible
 github.com/pingcap/tidb-tools v4.0.9-0.20201127090955-2707c97b3853+incompatible/go.mod h1:XGdcy9+yqlDSEMTpOXnwf3hiTeqrV6MN/u1se9N8yIM=
 github.com/pingcap/tipb v0.0.0-20190428032612-535e1abaa330/go.mod h1:RtkHW8WbcNxj8lsbzjaILci01CtYnYbIkQhjyZWrWVI=
 github.com/pingcap/tipb v0.0.0-20200417094153-7316d94df1ee/go.mod h1:RtkHW8WbcNxj8lsbzjaILci01CtYnYbIkQhjyZWrWVI=
-github.com/pingcap/tipb v0.0.0-20201209065231-aa39b1b86217 h1:Ophn4Ud/QHp1BH0FJOzbAVBW9Mw8BlX0gtWkK7ubDy0=
-github.com/pingcap/tipb v0.0.0-20201209065231-aa39b1b86217/go.mod h1:RtkHW8WbcNxj8lsbzjaILci01CtYnYbIkQhjyZWrWVI=
+github.com/pingcap/tipb v0.0.0-20201215091753-bd0cb2b314a4 h1:x64INZ8imEXO3MFcWD99lYlp52V9ZdYrxj74ynfyg3c=
+github.com/pingcap/tipb v0.0.0-20201215091753-bd0cb2b314a4/go.mod h1:RtkHW8WbcNxj8lsbzjaILci01CtYnYbIkQhjyZWrWVI=
 github.com/pingcap/tiup v1.2.3 h1:8OCQF7sHhT6VqE8pZU1JTSogPA90OFuWWM/B746x0YY=
 github.com/pingcap/tiup v1.2.3/go.mod h1:q8WzflNHjE1U49k2qstTL0clx2pKh8pkOzUFV4RTvQo=
 github.com/pkg/browser v0.0.0-20180916011732-0a3d74bf9ce4/go.mod h1:4OwLy04Bl9Ef3GJJCoec+30X3LQs/0/m4HFRt/2LUSA=

planner/core/planbuilder.go

@@ -3476,9 +3476,9 @@ func buildShowSchema(s *ast.ShowStmt, isView bool, isSequence bool) (schema *exp
 			mysql.TypeLonglong, mysql.TypeLonglong, mysql.TypeDouble, mysql.TypeDouble}
 	case ast.ShowStatsBuckets:
 		names = []string{"Db_name", "Table_name", "Partition_name", "Column_name", "Is_index", "Bucket_id", "Count",
-			"Repeats", "Lower_Bound", "Upper_Bound"}
+			"Repeats", "Lower_Bound", "Upper_Bound", "Ndv"}
 		ftypes = []byte{mysql.TypeVarchar, mysql.TypeVarchar, mysql.TypeVarchar, mysql.TypeVarchar, mysql.TypeTiny, mysql.TypeLonglong,
-			mysql.TypeLonglong, mysql.TypeLonglong, mysql.TypeVarchar, mysql.TypeVarchar}
+			mysql.TypeLonglong, mysql.TypeLonglong, mysql.TypeVarchar, mysql.TypeVarchar, mysql.TypeLonglong}
 	case ast.ShowStatsTopN:
 		names = []string{"Db_name", "Table_name", "Partition_name", "Column_name", "Is_index", "Value", "Count"}
 		ftypes = []byte{mysql.TypeVarchar, mysql.TypeVarchar, mysql.TypeVarchar, mysql.TypeVarchar, mysql.TypeTiny, mysql.TypeVarchar, mysql.TypeLonglong}

session/bootstrap.go

@@ -202,7 +202,8 @@ const (
 		count 		BIGINT(64) NOT NULL,
 		repeats 	BIGINT(64) NOT NULL,
 		upper_bound BLOB NOT NULL,
-		lower_bound BLOB ,
+		lower_bound BLOB , 
+		ndv         BIGINT NOT NULL DEFAULT 0,
 		UNIQUE INDEX tbl(table_id, is_index, hist_id, bucket_id)
 	);`
 
@@ -444,9 +445,11 @@ const (
 	version57 = 57
 	// version58 add `Repl_client_priv` and `Repl_slave_priv` to `mysql.user`
 	version58 = 58
+	// version59 add column ndv for mysql.stats_buckets.
+	version59 = 59
 
 	// please make sure this is the largest version
-	currentBootstrapVersion = version58
+	currentBootstrapVersion = version59
 )
 
 var (
@@ -509,6 +512,7 @@ var (
 		upgradeToVer56,
 		upgradeToVer57,
 		upgradeToVer58,
+		upgradeToVer59,
 	}
 )
 
@@ -1274,6 +1278,13 @@ func writeMemoryQuotaQuery(s Session) {
 	mustExecute(s, sql)
 }
 
+func upgradeToVer59(s Session, ver int64) {
+	if ver >= version59 {
+		return
+	}
+	doReentrantDDL(s, "ALTER TABLE mysql.stats_buckets ADD COLUMN `ndv` bigint not null default 0", infoschema.ErrColumnExists)
+}
+
 // updateBootstrapVer updates bootstrap version variable in mysql.TiDB table.
 func updateBootstrapVer(s Session) {
 	// Update bootstrap version.