Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

planner: a better way to round scale factor when collecting TopN stats #49808

Merged
merged 3 commits into from
Dec 27, 2023
Merged
Show file tree
Hide file tree
Changes from all commits
Commits
File filter

Filter by extension

Filter by extension

Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
2 changes: 1 addition & 1 deletion pkg/statistics/BUILD.bazel
Original file line number Diff line number Diff line change
Expand Up @@ -78,7 +78,7 @@ go_test(
data = glob(["testdata/**"]),
embed = [":statistics"],
flaky = True,
shard_count = 34,
shard_count = 35,
deps = [
"//pkg/config",
"//pkg/parser/ast",
Expand Down
4 changes: 1 addition & 3 deletions pkg/statistics/builder.go
Original file line number Diff line number Diff line change
Expand Up @@ -429,12 +429,10 @@ func BuildHistAndTopN(
}
}
}
for i := 0; i < len(topNList); i++ {
topNList[i].Count *= uint64(sampleFactor)
}
}

topn := &TopN{TopN: topNList}
topn.Scale(sampleFactor)

if uint64(count) <= topn.TotalCount() || int(hg.NDV) <= len(topn.TopN) {
// TopN includes all sample data
Expand Down
7 changes: 7 additions & 0 deletions pkg/statistics/cmsketch.go
Original file line number Diff line number Diff line change
Expand Up @@ -536,6 +536,13 @@ type TopN struct {
TopN []TopNMeta
}

// Scale scales the TopN by the given factor.
func (c *TopN) Scale(scaleFactor float64) {
for i := range c.TopN {
c.TopN[i].Count = uint64(float64(c.TopN[i].Count) * scaleFactor)
}
}

// AppendTopN appends a topn into the TopN struct.
func (c *TopN) AppendTopN(data []byte, count uint64) {
if c == nil {
Expand Down
20 changes: 20 additions & 0 deletions pkg/statistics/cmsketch_test.go
Original file line number Diff line number Diff line change
Expand Up @@ -266,3 +266,23 @@ func TestSortTopnMeta(t *testing.T) {
SortTopnMeta(data)
require.Equal(t, uint64(2), data[0].Count)
}

func TestTopNScale(t *testing.T) {
for _, scaleFactor := range []float64{0.9999, 1.00001, 1.9999, 4.9999, 5.001, 9.99} {
var data []TopNMeta
sumCount := uint64(0)
for i := 0; i < 20; i++ {
cnt := uint64(rand.Intn(100000))
data = append(data, TopNMeta{
Count: cnt,
})
sumCount += cnt
}
topN := TopN{TopN: data}
topN.Scale(scaleFactor)
scaleCount := float64(sumCount) * scaleFactor
delta := math.Abs(float64(topN.TotalCount()) - scaleCount)
roundErrorRatio := delta / scaleCount
require.Less(t, roundErrorRatio, 0.0001)
}
}