stats: fix estimation in index point query (#4959)

pingcap · Oct 31, 2017 · 943d77b · 943d77b
1 parent 08bd8c5
commit 943d77b
Show file tree

Hide file tree

Showing 2 changed files with 75 additions and 3 deletions.
diff --git a/statistics/histogram.go b/statistics/histogram.go
@@ -551,13 +551,23 @@ func (idx *Index) getRowCount(sc *variable.StatementContext, indexRanges []*type
 		if err != nil {
 			return 0, errors.Trace(err)
 		}
-		if indexRange.LowExclude {
-			lb = append(lb, 0)
-		}
 		rb, err := codec.EncodeKey(nil, indexRange.HighVal...)
 		if err != nil {
 			return 0, errors.Trace(err)
 		}
+		if string(lb) == string(rb) {
+			if !indexRange.LowExclude && !indexRange.HighExclude {
+				rowCount, err := idx.equalRowCount(sc, types.NewBytesDatum(lb))
+				if err != nil {
+					return 0, errors.Trace(err)
+				}
+				totalCount += rowCount
+			}
+			continue
+		}
+		if indexRange.LowExclude {
+			lb = append(lb, 0)
+		}
 		if !indexRange.HighExclude {
 			rb = append(rb, 0)
 		}

diff --git a/statistics/statistics_test.go b/statistics/statistics_test.go
@@ -516,3 +516,65 @@ func (s *testStatisticsSuite) TestIntColumnRanges(c *C) {
 	c.Assert(err, IsNil)
 	c.Assert(int(count), Equals, 1)
 }
+
+func (s *testStatisticsSuite) TestIndexRanges(c *C) {
+	bucketCount := int64(256)
+	ctx := mock.NewContext()
+	sc := ctx.GetSessionVars().StmtCtx
+
+	s.rc.(*recordSet).cursor = 0
+	rowCount, hg, err := BuildIndex(ctx, bucketCount, 0, s.rc)
+	calculateScalar(hg)
+	c.Check(err, IsNil)
+	c.Check(rowCount, Equals, int64(100000))
+	idxInfo := &model.IndexInfo{Columns: []*model.IndexColumn{{Offset: 0}}}
+	idx := &Index{Histogram: *hg, Info: idxInfo}
+	tbl := &Table{
+		Count:   int64(idx.totalRowCount()),
+		Indices: make(map[int64]*Index),
+	}
+	ran := []*types.IndexRange{{
+		LowVal:  []types.Datum{types.MinNotNullDatum()},
+		HighVal: []types.Datum{types.MaxValueDatum()},
+	}}
+	count, err := tbl.GetRowCountByIndexRanges(sc, 0, ran)
+	c.Assert(err, IsNil)
+	c.Assert(int(count), Equals, 99900)
+	ran[0].LowVal[0] = types.NewIntDatum(1000)
+	ran[0].HighVal[0] = types.NewIntDatum(2000)
+	count, err = tbl.GetRowCountByIndexRanges(sc, 0, ran)
+	c.Assert(err, IsNil)
+	c.Assert(int(count), Equals, 2500)
+	ran[0].LowVal[0] = types.NewIntDatum(1001)
+	ran[0].HighVal[0] = types.NewIntDatum(1999)
+	count, err = tbl.GetRowCountByIndexRanges(sc, 0, ran)
+	c.Assert(err, IsNil)
+	c.Assert(int(count), Equals, 2500)
+	ran[0].LowVal[0] = types.NewIntDatum(1000)
+	ran[0].HighVal[0] = types.NewIntDatum(1000)
+	count, err = tbl.GetRowCountByIndexRanges(sc, 0, ran)
+	c.Assert(err, IsNil)
+	c.Assert(int(count), Equals, 100)
+
+	tbl.Indices[0] = idx
+	ran[0].LowVal[0] = types.MinNotNullDatum()
+	ran[0].HighVal[0] = types.MaxValueDatum()
+	count, err = tbl.GetRowCountByIndexRanges(sc, 0, ran)
+	c.Assert(err, IsNil)
+	c.Assert(int(count), Equals, 100000)
+	ran[0].LowVal[0] = types.NewIntDatum(1000)
+	ran[0].HighVal[0] = types.NewIntDatum(2000)
+	count, err = tbl.GetRowCountByIndexRanges(sc, 0, ran)
+	c.Assert(err, IsNil)
+	c.Assert(int(count), Equals, 999)
+	ran[0].LowVal[0] = types.NewIntDatum(1001)
+	ran[0].HighVal[0] = types.NewIntDatum(1990)
+	count, err = tbl.GetRowCountByIndexRanges(sc, 0, ran)
+	c.Assert(err, IsNil)
+	c.Assert(int(count), Equals, 988)
+	ran[0].LowVal[0] = types.NewIntDatum(1000)
+	ran[0].HighVal[0] = types.NewIntDatum(1000)
+	count, err = tbl.GetRowCountByIndexRanges(sc, 0, ran)
+	c.Assert(err, IsNil)
+	c.Assert(int(count), Equals, 1)
+}