From 7ac87990d15f9ada641140a295f3c2b9f56de8af Mon Sep 17 00:00:00 2001
From: Jakub Martin <jakub.wit.martin@gmail.com>
Date: Mon, 10 Oct 2022 01:03:35 +0200
Subject: [PATCH] Hashing optimizations to reduce unnecessary allocations.

---
 aggregates/distinct.go             |  6 +---
 execution/nodes/distinct.go        |  7 +----
 execution/nodes/simple_group_by.go |  7 +----
 go.mod                             |  2 +-
 octosql/values.go                  | 50 +++++++++++++++++-------------
 5 files changed, 32 insertions(+), 40 deletions(-)

diff --git a/aggregates/distinct.go b/aggregates/distinct.go
index 47a25e84..03ae0929 100644
--- a/aggregates/distinct.go
+++ b/aggregates/distinct.go
@@ -1,8 +1,6 @@
 package aggregates
 
 import (
-	"hash/fnv"
-
 	"github.com/zyedidia/generic/hashmap"
 
 	"github.com/cube2222/octosql/execution"
@@ -37,9 +35,7 @@ func NewDistinctPrototype(wrapped func() nodes.Aggregate) func() nodes.Aggregate
 				func(a, b octosql.Value) bool {
 					return a.Compare(b) == 0
 				}, func(v octosql.Value) uint64 {
-					hash := fnv.New64()
-					v.Hash(hash)
-					return hash.Sum64()
+					return v.Hash()
 				}),
 			wrapped: wrapped(),
 		}
diff --git a/execution/nodes/distinct.go b/execution/nodes/distinct.go
index bde6ebad..6a3588c5 100644
--- a/execution/nodes/distinct.go
+++ b/execution/nodes/distinct.go
@@ -2,7 +2,6 @@ package nodes
 
 import (
 	"fmt"
-	"hash/fnv"
 
 	"github.com/zyedidia/generic/hashmap"
 
@@ -35,11 +34,7 @@ func (o *Distinct) Run(execCtx ExecutionContext, produce ProduceFn, metaSend Met
 			}
 			return true
 		}, func(k []octosql.Value) uint64 {
-			hash := fnv.New64()
-			for _, v := range k {
-				v.Hash(hash)
-			}
-			return hash.Sum64()
+			return octosql.HashManyValues(k)
 		})
 	o.source.Run(
 		execCtx,
diff --git a/execution/nodes/simple_group_by.go b/execution/nodes/simple_group_by.go
index 24328736..8f0cc809 100644
--- a/execution/nodes/simple_group_by.go
+++ b/execution/nodes/simple_group_by.go
@@ -2,7 +2,6 @@ package nodes
 
 import (
 	"fmt"
-	"hash/fnv"
 	"time"
 
 	"github.com/zyedidia/generic/hashmap"
@@ -52,11 +51,7 @@ func (g *SimpleGroupBy) Run(ctx ExecutionContext, produce ProduceFn, metaSend Me
 		}
 		return true
 	}, func(k GroupKey) uint64 {
-		hash := fnv.New64()
-		for _, v := range k {
-			v.Hash(hash)
-		}
-		return hash.Sum64()
+		return octosql.HashManyValues(k)
 	})
 
 	if err := g.source.Run(ctx, func(produceCtx ProduceContext, record Record) error {
diff --git a/go.mod b/go.mod
index 638b1214..06390d9f 100644
--- a/go.mod
+++ b/go.mod
@@ -20,6 +20,7 @@ require (
 	github.com/pkg/errors v0.9.1
 	github.com/pkg/profile v1.6.0
 	github.com/pmezard/go-difflib v1.0.0
+	github.com/segmentio/fasthash v1.0.3
 	github.com/segmentio/parquet-go v0.0.0-20220421002521-93f8e5ed3407
 	github.com/skratchdot/open-golang v0.0.0-20200116055534-eef842397966
 	github.com/spf13/cobra v1.4.0
@@ -59,7 +60,6 @@ require (
 	github.com/pkg/term v1.2.0-beta.2 // indirect
 	github.com/rivo/uniseg v0.2.0 // indirect
 	github.com/segmentio/encoding v0.3.5 // indirect
-	github.com/segmentio/fasthash v1.0.3 // indirect
 	github.com/shopspring/decimal v1.2.0 // indirect
 	github.com/spf13/pflag v1.0.5 // indirect
 	github.com/ulikunitz/xz v0.5.10 // indirect
diff --git a/octosql/values.go b/octosql/values.go
index 31700165..1e54c31d 100644
--- a/octosql/values.go
+++ b/octosql/values.go
@@ -1,12 +1,12 @@
 package octosql
 
 import (
-	"encoding/binary"
 	"fmt"
-	"hash"
 	"math"
 	"strings"
 	"time"
+
+	"github.com/segmentio/fasthash/fnv1a"
 )
 
 var ZeroValue = Value{}
@@ -239,54 +239,58 @@ func (value Value) Compare(other Value) int {
 	}
 }
 
-func (value Value) Hash(hash hash.Hash64) {
+func (value Value) Hash() uint64 {
+	return value.hash(fnv1a.Init64)
+}
+
+func HashManyValues(values []Value) uint64 {
+	hash := fnv1a.Init64
+	for _, v := range values {
+		hash = v.hash(hash)
+	}
+	return hash
+}
+
+func (value Value) hash(hash uint64) uint64 {
 	switch value.TypeID {
 	case TypeIDNull:
-		hash.Write([]byte{0})
+		hash = fnv1a.AddUint64(hash, 0)
 
 	case TypeIDInt:
-		var data [8]byte
-		binary.BigEndian.PutUint64(data[:], uint64(value.Int))
-		hash.Write(data[:])
+		hash = fnv1a.AddUint64(hash, uint64(value.Int))
 
 	case TypeIDFloat:
-		var data [8]byte
-		binary.BigEndian.PutUint64(data[:], math.Float64bits(value.Float))
-		hash.Write(data[:])
+		hash = fnv1a.AddUint64(hash, math.Float64bits(value.Float))
 
 	case TypeIDBoolean:
 		if value.Boolean {
-			hash.Write([]byte{1})
+			hash = fnv1a.AddUint64(hash, 1)
 		} else {
-			hash.Write([]byte{0})
+			hash = fnv1a.AddUint64(hash, 0)
 		}
 
 	case TypeIDString:
-		hash.Write([]byte(value.Str))
+		hash = fnv1a.AddString64(hash, value.Str)
 
 	case TypeIDTime:
-		var data [8]byte
-		binary.BigEndian.PutUint64(data[:], uint64(value.Time.UnixNano()))
-		hash.Write(data[:])
+		hash = fnv1a.AddUint64(hash, uint64(value.Time.UnixNano()))
 
 	case TypeIDDuration:
-		var data [8]byte
-		binary.BigEndian.PutUint64(data[:], uint64(value.Duration))
-		hash.Write(data[:])
+		hash = fnv1a.AddUint64(hash, uint64(value.Duration))
 
 	case TypeIDList:
 		for i := range value.List {
-			value.List[i].Hash(hash)
+			hash = value.List[i].hash(hash)
 		}
 
 	case TypeIDStruct:
 		for i := range value.List {
-			value.Struct[i].Hash(hash)
+			hash = value.Struct[i].hash(hash)
 		}
 
 	case TypeIDTuple:
 		for i := range value.List {
-			value.Tuple[i].Hash(hash)
+			hash = value.Tuple[i].hash(hash)
 		}
 
 	case TypeIDUnion:
@@ -294,6 +298,8 @@ func (value Value) Hash(hash hash.Hash64) {
 	default:
 		panic("impossible, type switch bug")
 	}
+
+	return hash
 }
 
 func (value Value) Equal(other Value) bool {