mirror of https://github.com/jackc/pgx.git
Improve stdlib performance with large result sets
In a hot path type assertions are expensive. Cache the already correctly
typed interfaces. ~20% improvement with 1000 rows.
Before:
jack@glados ~/dev/pgx/stdlib ±master » PGX_BENCH_SELECT_ROWS_COUNTS='1 10 1000' got ./... -bench=SelectRows -benchmem
goos: darwin
goarch: amd64
pkg: github.com/jackc/pgx/v4/stdlib
BenchmarkSelectRowsScanSimple/1_rows-16 21465 55060 ns/op 1679 B/op 40 allocs/op
BenchmarkSelectRowsScanSimple/10_rows-16 16692 71176 ns/op 3827 B/op 148 allocs/op
BenchmarkSelectRowsScanSimple/1000_rows-16 800 1369547 ns/op 248855 B/op 12938 allocs/op
BenchmarkSelectRowsScanNull/1_rows-16 20306 57883 ns/op 1940 B/op 54 allocs/op
BenchmarkSelectRowsScanNull/10_rows-16 15942 74729 ns/op 4294 B/op 171 allocs/op
BenchmarkSelectRowsScanNull/1000_rows-16 829 1326788 ns/op 261291 B/op 13051 allocs/op
PASS
ok github.com/jackc/pgx/v4/stdlib 10.429s
After:
jack@glados ~/dev/pgx/stdlib ±master⚡ » PGX_BENCH_SELECT_ROWS_COUNTS='1 10 1000' got ./... -bench=SelectRows -benchmem
goos: darwin
goarch: amd64
pkg: github.com/jackc/pgx/v4/stdlib
BenchmarkSelectRowsScanSimple/1_rows-16 21327 55097 ns/op 2127 B/op 43 allocs/op
BenchmarkSelectRowsScanSimple/10_rows-16 16724 69496 ns/op 4276 B/op 151 allocs/op
BenchmarkSelectRowsScanSimple/1000_rows-16 1009 1124573 ns/op 250037 B/op 12941 allocs/op
BenchmarkSelectRowsScanNull/1_rows-16 20577 58117 ns/op 2396 B/op 57 allocs/op
BenchmarkSelectRowsScanNull/10_rows-16 16402 72533 ns/op 4750 B/op 174 allocs/op
BenchmarkSelectRowsScanNull/1000_rows-16 1010 1161437 ns/op 261735 B/op 13054 allocs/op
PASS
ok github.com/jackc/pgx/v4/stdlib 10.363s
pull/745/head
Jack Christensen
parent
dea915e605
commit
c03ac1519e
128
stdlib/sql.go
128
stdlib/sql.go
|
|
@ -352,6 +352,9 @@ type Rows struct {
|
|||
conn *Conn
|
||||
rows pgx.Rows
|
||||
values []interface{}
|
||||
driverValuers []driver.Valuer
|
||||
textDecoders []pgtype.TextDecoder
|
||||
binaryDecoders []pgtype.BinaryDecoder
|
||||
skipNext bool
|
||||
skipNextMore bool
|
||||
}
|
||||
|
|
@ -444,42 +447,112 @@ func (r *Rows) Close() error {
|
|||
}
|
||||
|
||||
func (r *Rows) Next(dest []driver.Value) error {
|
||||
ci := r.conn.conn.ConnInfo()
|
||||
fieldDescriptions := r.rows.FieldDescriptions()
|
||||
|
||||
if r.values == nil {
|
||||
r.values = make([]interface{}, len(r.rows.FieldDescriptions()))
|
||||
for i, fd := range r.rows.FieldDescriptions() {
|
||||
r.values = make([]interface{}, len(fieldDescriptions))
|
||||
r.driverValuers = make([]driver.Valuer, len(fieldDescriptions))
|
||||
r.textDecoders = make([]pgtype.TextDecoder, len(fieldDescriptions))
|
||||
r.binaryDecoders = make([]pgtype.BinaryDecoder, len(fieldDescriptions))
|
||||
|
||||
for i, fd := range fieldDescriptions {
|
||||
switch fd.DataTypeOID {
|
||||
case pgtype.BoolOID:
|
||||
r.values[i] = &pgtype.Bool{}
|
||||
v := &pgtype.Bool{}
|
||||
r.values[i] = v
|
||||
r.driverValuers[i] = v
|
||||
r.textDecoders[i] = v
|
||||
r.binaryDecoders[i] = v
|
||||
case pgtype.ByteaOID:
|
||||
r.values[i] = &pgtype.Bytea{}
|
||||
v := &pgtype.Bytea{}
|
||||
r.values[i] = v
|
||||
r.driverValuers[i] = v
|
||||
r.textDecoders[i] = v
|
||||
r.binaryDecoders[i] = v
|
||||
case pgtype.CIDOID:
|
||||
r.values[i] = &pgtype.CID{}
|
||||
v := &pgtype.CID{}
|
||||
r.values[i] = v
|
||||
r.driverValuers[i] = v
|
||||
r.textDecoders[i] = v
|
||||
r.binaryDecoders[i] = v
|
||||
case pgtype.DateOID:
|
||||
r.values[i] = &pgtype.Date{}
|
||||
v := &pgtype.Date{}
|
||||
r.values[i] = v
|
||||
r.driverValuers[i] = v
|
||||
r.textDecoders[i] = v
|
||||
r.binaryDecoders[i] = v
|
||||
case pgtype.Float4OID:
|
||||
r.values[i] = &pgtype.Float4{}
|
||||
v := &pgtype.Float4{}
|
||||
r.values[i] = v
|
||||
r.driverValuers[i] = v
|
||||
r.textDecoders[i] = v
|
||||
r.binaryDecoders[i] = v
|
||||
case pgtype.Float8OID:
|
||||
r.values[i] = &pgtype.Float8{}
|
||||
v := &pgtype.Float8{}
|
||||
r.values[i] = v
|
||||
r.driverValuers[i] = v
|
||||
r.textDecoders[i] = v
|
||||
r.binaryDecoders[i] = v
|
||||
case pgtype.Int2OID:
|
||||
r.values[i] = &pgtype.Int2{}
|
||||
v := &pgtype.Int2{}
|
||||
r.values[i] = v
|
||||
r.driverValuers[i] = v
|
||||
r.textDecoders[i] = v
|
||||
r.binaryDecoders[i] = v
|
||||
case pgtype.Int4OID:
|
||||
r.values[i] = &pgtype.Int4{}
|
||||
v := &pgtype.Int4{}
|
||||
r.values[i] = v
|
||||
r.driverValuers[i] = v
|
||||
r.textDecoders[i] = v
|
||||
r.binaryDecoders[i] = v
|
||||
case pgtype.Int8OID:
|
||||
r.values[i] = &pgtype.Int8{}
|
||||
v := &pgtype.Int8{}
|
||||
r.values[i] = v
|
||||
r.driverValuers[i] = v
|
||||
r.textDecoders[i] = v
|
||||
r.binaryDecoders[i] = v
|
||||
case pgtype.JSONOID:
|
||||
r.values[i] = &pgtype.JSON{}
|
||||
v := &pgtype.JSON{}
|
||||
r.values[i] = v
|
||||
r.driverValuers[i] = v
|
||||
r.textDecoders[i] = v
|
||||
r.binaryDecoders[i] = v
|
||||
case pgtype.JSONBOID:
|
||||
r.values[i] = &pgtype.JSONB{}
|
||||
v := &pgtype.JSONB{}
|
||||
r.values[i] = v
|
||||
r.driverValuers[i] = v
|
||||
r.textDecoders[i] = v
|
||||
r.binaryDecoders[i] = v
|
||||
case pgtype.OIDOID:
|
||||
r.values[i] = &pgtype.OIDValue{}
|
||||
v := &pgtype.OIDValue{}
|
||||
r.values[i] = v
|
||||
r.driverValuers[i] = v
|
||||
r.textDecoders[i] = v
|
||||
r.binaryDecoders[i] = v
|
||||
case pgtype.TimestampOID:
|
||||
r.values[i] = &pgtype.Timestamp{}
|
||||
v := &pgtype.Timestamp{}
|
||||
r.values[i] = v
|
||||
r.driverValuers[i] = v
|
||||
r.textDecoders[i] = v
|
||||
r.binaryDecoders[i] = v
|
||||
case pgtype.TimestamptzOID:
|
||||
r.values[i] = &pgtype.Timestamptz{}
|
||||
v := &pgtype.Timestamptz{}
|
||||
r.values[i] = v
|
||||
r.driverValuers[i] = v
|
||||
r.textDecoders[i] = v
|
||||
r.binaryDecoders[i] = v
|
||||
case pgtype.XIDOID:
|
||||
r.values[i] = &pgtype.XID{}
|
||||
v := &pgtype.XID{}
|
||||
r.values[i] = v
|
||||
r.driverValuers[i] = v
|
||||
r.textDecoders[i] = v
|
||||
r.binaryDecoders[i] = v
|
||||
default:
|
||||
r.values[i] = &pgtype.GenericText{}
|
||||
v := &pgtype.GenericText{}
|
||||
r.values[i] = v
|
||||
r.driverValuers[i] = v
|
||||
r.textDecoders[i] = v
|
||||
}
|
||||
}
|
||||
}
|
||||
|
|
@ -500,15 +573,24 @@ func (r *Rows) Next(dest []driver.Value) error {
|
|||
}
|
||||
}
|
||||
|
||||
err := r.rows.Scan(r.values...)
|
||||
for i, rv := range r.rows.RawValues() {
|
||||
fd := fieldDescriptions[i]
|
||||
if fd.Format == pgx.BinaryFormatCode {
|
||||
err := r.binaryDecoders[i].DecodeBinary(ci, rv)
|
||||
if err != nil {
|
||||
return err
|
||||
return fmt.Errorf("scan field %d failed: %v", i, err)
|
||||
}
|
||||
} else {
|
||||
err := r.textDecoders[i].DecodeText(ci, rv)
|
||||
if err != nil {
|
||||
return fmt.Errorf("scan field %d failed: %v", i, err)
|
||||
}
|
||||
}
|
||||
|
||||
for i, v := range r.values {
|
||||
dest[i], err = v.(driver.Valuer).Value()
|
||||
var err error
|
||||
dest[i], err = r.driverValuers[i].Value()
|
||||
if err != nil {
|
||||
return err
|
||||
return fmt.Errorf("convert field %d failed: %v", i, err)
|
||||
}
|
||||
}
|
||||
|
||||
|
|
|
|||
Loading…
Reference in New Issue