Merge pull request #7497 from influxdata/js-6704-first-last-optimization

Optimize first/last when no group by interval is present

CHANGELOG.md

@@ -39,6 +39,7 @@ The following configuration changes in the `[data]` section may need to changed
 - [#7480](https://github.com/influxdata/influxdb/pull/7480): Improve compaction planning performance by caching tsm file stats.
 - [#7320](https://github.com/influxdata/influxdb/issues/7320): Update defaults in config for latest best practices
 - [#7495](https://github.com/influxdata/influxdb/pull/7495): Rewrite regexes of the form host = /^server-a$/ to host = 'server-a', to take advantage of the tsdb index.
+- [#6704](https://github.com/influxdata/influxdb/issues/6704): Optimize first/last when no group by interval is present.
 
 ### Bugfixes
 

tsdb/engine/tsm1/engine.go

@@ -1222,7 +1222,22 @@ func (e *Engine) CreateIterator(opt influxql.IteratorOptions) (influxql.Iterator
 	if call, ok := opt.Expr.(*influxql.Call); ok {
 		refOpt := opt
 		refOpt.Expr = call.Args[0].(*influxql.VarRef)
-		inputs, err := e.createVarRefIterator(refOpt, true)
+
+		aggregate := true
+		if opt.Interval.IsZero() {
+			switch call.Name {
+			case "first":
+				aggregate = false
+				refOpt.Limit = 1
+				refOpt.Ascending = true
+			case "last":
+				aggregate = false
+				refOpt.Limit = 1
+				refOpt.Ascending = false
+			}
+		}
+
+		inputs, err := e.createVarRefIterator(refOpt, aggregate)
 		if err != nil {
 			return nil, err
 		} else if len(inputs) == 0 {

tsdb/engine/tsm1/engine_test.go

@@ -629,6 +629,48 @@ func benchmarkEngineCreateIteratorCount(b *testing.B, pointN int) {
 	}, pointN)
 }
 
+func BenchmarkEngine_CreateIterator_First_1K(b *testing.B) {
+	benchmarkEngineCreateIteratorFirst(b, 1000)
+}
+func BenchmarkEngine_CreateIterator_First_100K(b *testing.B) {
+	benchmarkEngineCreateIteratorFirst(b, 100000)
+}
+func BenchmarkEngine_CreateIterator_First_1M(b *testing.B) {
+	benchmarkEngineCreateIteratorFirst(b, 1000000)
+}
+
+func benchmarkEngineCreateIteratorFirst(b *testing.B, pointN int) {
+	benchmarkIterator(b, influxql.IteratorOptions{
+		Expr:       influxql.MustParseExpr("first(value)"),
+		Sources:    []influxql.Source{&influxql.Measurement{Name: "cpu"}},
+		Dimensions: []string{"host"},
+		Ascending:  true,
+		StartTime:  influxql.MinTime,
+		EndTime:    influxql.MaxTime,
+	}, pointN)
+}
+
+func BenchmarkEngine_CreateIterator_Last_1K(b *testing.B) {
+	benchmarkEngineCreateIteratorLast(b, 1000)
+}
+func BenchmarkEngine_CreateIterator_Last_100K(b *testing.B) {
+	benchmarkEngineCreateIteratorLast(b, 100000)
+}
+func BenchmarkEngine_CreateIterator_Last_1M(b *testing.B) {
+	benchmarkEngineCreateIteratorLast(b, 1000000)
+}
+
+func benchmarkEngineCreateIteratorLast(b *testing.B, pointN int) {
+	benchmarkIterator(b, influxql.IteratorOptions{
+		Expr:       influxql.MustParseExpr("last(value)"),
+		Sources:    []influxql.Source{&influxql.Measurement{Name: "cpu"}},
+		Dimensions: []string{"host"},
+		Ascending:  true,
+		StartTime:  influxql.MinTime,
+		EndTime:    influxql.MaxTime,
+	}, pointN)
+}
+
 func BenchmarkEngine_CreateIterator_Limit_1K(b *testing.B) {
 	benchmarkEngineCreateIteratorLimit(b, 1000)
 }