From 7fb015cc7f88391f9e7ffcc3ad492b1b09d0d536 Mon Sep 17 00:00:00 2001
From: Stuart Carnie <stuart.carnie@gmail.com>
Date: Tue, 20 Nov 2018 11:55:04 -0700
Subject: [PATCH] chore: Add GroupBy benchmark

---
 pkg/testing/gen/sequence.go           | 59 +++++++++++++++++
 pkg/testing/gen/tags_sequence.go      | 93 +++++++++++++++++++++++++++
 storage/reads/group_resultset_test.go | 33 ++++++++++
 3 files changed, 185 insertions(+)
 create mode 100644 pkg/testing/gen/sequence.go
 create mode 100644 pkg/testing/gen/tags_sequence.go

diff --git a/pkg/testing/gen/sequence.go b/pkg/testing/gen/sequence.go
new file mode 100644
index 0000000000..3942998821
--- /dev/null
+++ b/pkg/testing/gen/sequence.go
@@ -0,0 +1,59 @@
+package gen
+
+import (
+	"fmt"
+	"math"
+)
+
+type Sequence interface {
+	Next() bool
+	Value() string
+	Count() int
+}
+
+type CounterByteSequence struct {
+	format string
+	nfmt   string
+	val    string
+	s      int
+	v      int
+	end    int
+}
+
+func NewCounterByteSequenceCount(n int) *CounterByteSequence {
+	return NewCounterByteSequence("value%s", 0, n)
+}
+
+func NewCounterByteSequence(format string, start, end int) *CounterByteSequence {
+	s := &CounterByteSequence{
+		format: format,
+		nfmt:   fmt.Sprintf("%%0%dd", int(math.Ceil(math.Log10(float64(end))))),
+		s:      start,
+		v:      start,
+		end:    end,
+	}
+	s.update()
+	return s
+}
+
+func (s *CounterByteSequence) Next() bool {
+	s.v++
+	if s.v >= s.end {
+		s.v = s.s
+	}
+	s.update()
+	return true
+}
+
+func (s *CounterByteSequence) update() {
+	s.val = fmt.Sprintf(s.format, fmt.Sprintf(s.nfmt, s.v))
+}
+
+func (s *CounterByteSequence) Count() int    { return s.end - s.s }
+func (s *CounterByteSequence) Value() string { return s.val }
+
+type ConstantStringSequence string
+
+func (ConstantStringSequence) Next() bool      { return true }
+func (s ConstantStringSequence) Value() string { return string(s) }
+func (ConstantStringSequence) Count() int      { return 1 }
diff --git a/pkg/testing/gen/tags_sequence.go b/pkg/testing/gen/tags_sequence.go
new file mode 100644
index 0000000000..26b7889842
--- /dev/null
+++ b/pkg/testing/gen/tags_sequence.go
@@ -0,0 +1,93 @@
+package gen
+
+import (
+	"fmt"
+	"math"
+	"sort"
+
+	"github.com/influxdata/influxdb/models"
+)
+
+type TagsSequence interface {
+	Next() bool
+	Value() models.Tags
+	Count() int
+}
+
+type TagsValuesSequence struct {
+	tags models.Tags
+	vals []Sequence
+	n    int
+	max  int
+}
+
+func NewTagsValuesSequenceKeysValues(keys []string, vals []Sequence) *TagsValuesSequence {
+	tm := make(map[string]string, len(keys))
+	for _, k := range keys {
+		tm[k] = ""
+	}
+
+	count := 1
+	for i := range vals {
+		count *= vals[i].Count()
+	}
+
+	// models.Tags are ordered, so ensure vals are ordered with respect to keys
+	sort.Sort(keyValues{keys, vals})
+
+	return &TagsValuesSequence{
+		tags: models.NewTags(tm),
+		vals: vals,
+		max:  count,
+	}
+}
+
+func NewTagsValuesSequenceValues(prefix string, vals []Sequence) *TagsValuesSequence {
+	keys := make([]string, len(vals))
+	// max tag width
+	tw := int(math.Ceil(math.Log10(float64(len(vals)))))
+	tf := fmt.Sprintf("%s%%0%dd", prefix, tw)
+	for i := range vals {
+		keys[i] = fmt.Sprintf(tf, i)
+	}
+	return NewTagsValuesSequenceKeysValues(keys, vals)
+}
+
+func (s *TagsValuesSequence) Next() bool {
+	if s.n >= s.max {
+		return false
+	}
+
+	for i := range s.vals {
+		s.tags[i].Value = []byte(s.vals[i].Value())
+	}
+
+	s.n++
+	i := s.n
+	for j := len(s.vals) - 1; j >= 0; j-- {
+		v := s.vals[j]
+		v.Next()
+		c := v.Count()
+		if r := i % c; r != 0 {
+			break
+		}
+		i /= c
+	}
+
+	return true
+}
+
+func (s *TagsValuesSequence) Value() models.Tags { return s.tags }
+func (s *TagsValuesSequence) Count() int         { return s.max }
+
+type keyValues struct {
+	keys []string
+	vals []Sequence
+}
+
+func (k keyValues) Len() int           { return len(k.keys) }
+func (k keyValues) Less(i, j int) bool { return k.keys[i] < k.keys[j] }
+func (k keyValues) Swap(i, j int) {
+	k.keys[i], k.keys[j] = k.keys[j], k.keys[i]
+	k.vals[i], k.vals[j] = k.vals[j], k.vals[i]
+}
diff --git a/storage/reads/group_resultset_test.go b/storage/reads/group_resultset_test.go
index 2aedee8e72..d8a6c59103 100644
--- a/storage/reads/group_resultset_test.go
+++ b/storage/reads/group_resultset_test.go
@@ -307,3 +307,36 @@ func (s *sliceSeriesCursor) Next() *reads.SeriesRow {
 	}
 	return nil
 }
+
+func BenchmarkNewGroupResultSet_GroupBy(b *testing.B) {
+	card := []int{10, 10, 10}
+	vals := make([]gen.Sequence, len(card))
+	for i := range card {
+		vals[i] = gen.NewCounterByteSequenceCount(card[i])
+	}
+
+	tags := gen.NewTagsValuesSequenceValues("tag", vals)
+	rows := make([]reads.SeriesRow, tags.Count())
+	for i := range rows {
+		tags.Next()
+		t := tags.Value().Clone()
+		rows[i].SeriesTags = t
+		rows[i].Tags = t
+		rows[i].Name = []byte("m0")
+	}
+
+	cur := &sliceSeriesCursor{rows: rows}
+	newCursor := func() (reads.SeriesCursor, error) {
+		cur.i = 0
+		return cur, nil
+	}
+
+	b.ResetTimer()
+	b.ReportAllocs()
+	for i := 0; i < b.N; i++ {
+		var hints datatypes.HintFlags
+		hints.SetHintSchemaAllTime()
+		rs := reads.NewGroupResultSet(context.Background(), &datatypes.ReadRequest{Group: datatypes.GroupBy, GroupKeys: []string{"tag2"}, Hints: hints}, newCursor)
+		rs.Close()
+	}
+}