statistics: trigger auto-analyze based on histogram row count

pingcap · May 8, 2021 · 14cb4a7 · 14cb4a7
1 parent 04da3ce
commit 14cb4a7
Show file tree

Hide file tree

Showing 3 changed files with 60 additions and 2 deletions.
diff --git a/statistics/handle/update.go b/statistics/handle/update.go
@@ -875,10 +875,14 @@ func NeedAnalyzeTable(tbl *statistics.Table, limit time.Duration, autoAnalyzeRat
 		return false, ""
 	}
 	// No need to analyze it.
-	if float64(tbl.ModifyCount)/float64(tbl.Count) <= autoAnalyzeRatio {
+	tblCnt := float64(tbl.Count)
+	if histCnt := tbl.ColHistCount(); histCnt > 0 {
+		tblCnt = histCnt
+	}
+	if float64(tbl.ModifyCount)/tblCnt <= autoAnalyzeRatio {
 		return false, ""
 	}
-	return true, fmt.Sprintf("too many modifications(%v/%v>%v)", tbl.ModifyCount, tbl.Count, autoAnalyzeRatio)
+	return true, fmt.Sprintf("too many modifications(%v/%v>%v)", tbl.ModifyCount, tblCnt, autoAnalyzeRatio)
 }
 
 func (h *Handle) getAutoAnalyzeParameters() map[string]string {

diff --git a/statistics/handle/update_test.go b/statistics/handle/update_test.go
@@ -2257,3 +2257,47 @@ func (s *testSerialStatsSuite) TestAutoUpdatePartitionInDynamicOnlyMode(c *C) {
 		c.Assert(partitionStats.ModifyCount, Equals, int64(0))
 	})
 }
+
+func (s *testSerialStatsSuite) TestAutoAnalyzeRatio(c *C) {
+	defer cleanEnv(c, s.store, s.do)
+	tk := testkit.NewTestKit(c, s.store)
+
+	oriStart := tk.MustQuery("select @@tidb_auto_analyze_start_time").Rows()[0][0].(string)
+	oriEnd := tk.MustQuery("select @@tidb_auto_analyze_end_time").Rows()[0][0].(string)
+	handle.AutoAnalyzeMinCnt = 0
+	defer func() {
+		handle.AutoAnalyzeMinCnt = 1000
+		tk.MustExec(fmt.Sprintf("set global tidb_auto_analyze_start_time='%v'", oriStart))
+		tk.MustExec(fmt.Sprintf("set global tidb_auto_analyze_end_time='%v'", oriEnd))
+	}()
+
+	h := s.do.StatsHandle()
+	tk.MustExec("use test")
+	tk.MustExec("create table t (a int)")
+	c.Assert(h.HandleDDLEvent(<-h.DDLEventCh()), IsNil)
+	tk.MustExec("insert into t values (1)" + strings.Repeat(", (1)", 19))
+	c.Assert(h.DumpStatsDeltaToKV(handle.DumpAll), IsNil)
+	is := s.do.InfoSchema()
+	c.Assert(h.Update(is), IsNil)
+	// To pass the stats.Pseudo check in autoAnalyzeTable
+	tk.MustExec("analyze table t")
+	tk.MustExec("explain select * from t where a = 1")
+	c.Assert(h.LoadNeededHistograms(), IsNil)
+	tk.MustExec(fmt.Sprintf("set global tidb_auto_analyze_start_time='00:00 +0000'"))
+	tk.MustExec(fmt.Sprintf("set global tidb_auto_analyze_end_time='23:59 +0000'"))
+
+	tk.MustExec("insert into t values (1)" + strings.Repeat(", (1)", 10))
+	c.Assert(h.DumpStatsDeltaToKV(handle.DumpAll), IsNil)
+	c.Assert(h.Update(is), IsNil)
+	c.Assert(h.HandleAutoAnalyze(is), IsTrue)
+
+	tk.MustExec("delete from t limit 12")
+	c.Assert(h.DumpStatsDeltaToKV(handle.DumpAll), IsNil)
+	c.Assert(h.Update(is), IsNil)
+	c.Assert(h.HandleAutoAnalyze(is), IsFalse)
+
+	tk.MustExec("delete from t limit 4")
+	c.Assert(s.do.StatsHandle().DumpStatsDeltaToKV(handle.DumpAll), IsNil)
+	c.Assert(h.Update(s.do.InfoSchema()), IsNil)
+	c.Assert(s.do.StatsHandle().HandleAutoAnalyze(s.do.InfoSchema()), IsTrue)
+}
diff --git a/statistics/table.go b/statistics/table.go
@@ -214,6 +214,16 @@ func (t *Table) GetStatsInfo(ID int64, isIndex bool) (int64, *Histogram, *CMSket
 	return int64(colStatsInfo.TotalRowCount()), colStatsInfo.Histogram.Copy(), colStatsInfo.CMSketch.Copy(), colStatsInfo.TopN.Copy(), colStatsInfo.FMSketch.Copy()
 }
 
+// ColHistCount returns the count of the column histograms.
+func (t *Table) ColHistCount() float64 {
+	for _, col := range t.Columns {
+		if col != nil {
+			return col.TotalRowCount()
+		}
+	}
+	return -1
+}
+
 type tableColumnID struct {
 	TableID  int64
 	ColumnID int64