sql: avoid string to byte conversion copies on insert path

pkg/bench/bench_test.go

@@ -304,6 +304,36 @@ func runBenchmarkInsert(b *testing.B, db *sqlutils.SQLRunner, count int) {
 
 }
 
+// runBenchmarkInsertLarge benchmarks inserting count large rows into a table
+// where large means rows with a 1k string and 1k BYTES object.
+func runBenchmarkInsertLarge(b *testing.B, db *sqlutils.SQLRunner, count int) {
+	defer func() {
+		db.Exec(b, `DROP TABLE IF EXISTS bench.insert`)
+	}()
+
+	db.Exec(b, `CREATE TABLE bench.insert (k INT PRIMARY KEY, s STRING, b BYTES)`)
+	bigstr := strings.Repeat("x", 1<<10)
+	bigbytes := bytes.Repeat([]byte("x"), 1<<10)
+
+	b.ResetTimer()
+	var buf bytes.Buffer
+	val := 0
+	for i := 0; i < b.N; i++ {
+		buf.Reset()
+		buf.WriteString(`INSERT INTO bench.insert VALUES `)
+		for j := 0; j < count; j++ {
+			if j > 0 {
+				buf.WriteString(", ")
+			}
+			fmt.Fprintf(&buf, "(%d, '%s', '%s')", val, bigstr, bigbytes)
+			val++
+		}
+		db.Exec(b, buf.String())
+	}
+	b.StopTimer()
+
+}
+
 // runBenchmarkInsertFK benchmarks inserting count rows into a table with a
 // present foreign key into another table.
 func runBenchmarkInsertFK(b *testing.B, db *sqlutils.SQLRunner, count int) {
@@ -419,6 +449,7 @@ func BenchmarkSQL(b *testing.B) {
 		for _, runFn := range []func(*testing.B, *sqlutils.SQLRunner, int){
 			runBenchmarkDelete,
 			runBenchmarkInsert,
+			runBenchmarkInsertLarge,
 			runBenchmarkInsertDistinct,
 			runBenchmarkInsertFK,
 			runBenchmarkInsertSecondaryIndex,

pkg/sql/colenc/inverted.go

@@ -20,6 +20,7 @@ import (
 	"github.com/cockroachdb/cockroach/pkg/sql/sem/catid"
 	"github.com/cockroachdb/cockroach/pkg/sql/sem/tree"
 	"github.com/cockroachdb/cockroach/pkg/sql/types"
+	"github.com/cockroachdb/cockroach/pkg/util/encoding"
 	"github.com/cockroachdb/errors"
 )
 
@@ -32,7 +33,7 @@ func invertedColToDatum(vec coldata.Vec, row int) tree.Datum {
 		return tree.NewDJSON(vec.JSON().Get(row))
 	case types.StringFamily:
 		b := vec.Bytes().Get(row)
-		s := unsafeConvertBytesToString(b)
+		s := encoding.UnsafeConvertBytesToString(b)
 		return tree.NewDString(s)
 	}
 	// This handles arrays, geo etc.

pkg/sql/colenc/key.go

@@ -11,8 +11,6 @@
 package colenc
 
 import (
-	"unsafe"
-
 	"github.com/cockroachdb/cockroach/pkg/col/coldata"
 	"github.com/cockroachdb/cockroach/pkg/col/typeconv"
 	"github.com/cockroachdb/cockroach/pkg/roachpb"
@@ -144,9 +142,9 @@ func encodeKeys[T []byte | roachpb.Key](
 			}
 			s := ss.Get(r + start)
 			if dir == encoding.Ascending {
-				kys[r] = encoding.EncodeStringAscending(b, unsafeConvertBytesToString(s))
+				kys[r] = encoding.EncodeStringAscending(b, encoding.UnsafeConvertBytesToString(s))
 			} else {
-				kys[r] = encoding.EncodeStringDescending(b, unsafeConvertBytesToString(s))
+				kys[r] = encoding.EncodeStringDescending(b, encoding.UnsafeConvertBytesToString(s))
 			}
 		}
 	case types.TimestampFamily, types.TimestampTZFamily:
@@ -249,7 +247,3 @@ func (b *BatchEncoder) encodeIndexKey(
 	}
 	return nil
 }
-
-func unsafeConvertBytesToString(b []byte) string {
-	return *(*string)(unsafe.Pointer(&b))
-}

pkg/sql/colenc/legacy.go

@@ -16,6 +16,7 @@ import (
 	"github.com/cockroachdb/cockroach/pkg/sql/rowenc/valueside"
 	"github.com/cockroachdb/cockroach/pkg/sql/sem/tree"
 	"github.com/cockroachdb/cockroach/pkg/sql/types"
+	"github.com/cockroachdb/cockroach/pkg/util/encoding"
 	"github.com/cockroachdb/cockroach/pkg/util/json"
 	"github.com/cockroachdb/errors"
 )
@@ -66,7 +67,7 @@ func MarshalLegacy(colType *types.T, vec coldata.Vec, row int) (roachpb.Value, e
 		switch vec.Type().Family() {
 		case types.StringFamily, types.BytesFamily, types.UuidFamily, types.EnumFamily:
 			b := vec.Bytes().Get(row)
-			r.SetString(unsafeConvertBytesToString(b))
+			r.SetString(encoding.UnsafeConvertBytesToString(b))
 			return r, nil
 		}
 	case types.TimestampFamily, types.TimestampTZFamily:

pkg/sql/colexec/colexecproj/BUILD.bazel

@@ -30,6 +30,7 @@ go_library(
         "//pkg/sql/sqltelemetry",  # keep
         "//pkg/sql/types",  # keep
         "//pkg/util/duration",  # keep
+        "//pkg/util/encoding",  # keep
         "//pkg/util/json",  # keep
         "@com_github_cockroachdb_apd_v3//:apd",  # keep
         "@com_github_cockroachdb_errors//:errors",  # keep

pkg/sql/colexec/colexecproj/proj_non_const_ops_tmpl.go

@@ -40,6 +40,7 @@ import (
 	"github.com/cockroachdb/cockroach/pkg/sql/sqltelemetry"
 	"github.com/cockroachdb/cockroach/pkg/sql/types"
 	"github.com/cockroachdb/cockroach/pkg/util/duration"
+	"github.com/cockroachdb/cockroach/pkg/util/encoding"
 	"github.com/cockroachdb/errors"
 )
 

pkg/sql/colexec/colexecprojconst/BUILD.bazel

@@ -34,6 +34,7 @@ go_library(
         "//pkg/sql/sqltelemetry",  # keep
         "//pkg/sql/types",
         "//pkg/util/duration",  # keep
+        "//pkg/util/encoding",  # keep
         "//pkg/util/json",  # keep
         "@com_github_cockroachdb_apd_v3//:apd",  # keep
         "@com_github_cockroachdb_errors//:errors",