Skip to content

Commit 4fb2899

Browse files
charlieviethcharlievieth
committed
reduce CGO calls when scanning rows
This commit improves the performance of queries by at least 20% by only collecting the information needed to convert sqlite3 values to Go values and by batching CGO calls when scanning rows (the performance improvement scales with the number of columns being scanned). This commit adds a new coltype field to the SQLiteRows struct which stores the declared column type (either data/time or boolean) and the sqlite3 datatype. Previously, this library would fetch the string representation of each column, which is inefficient and rarely needed since the non-standard SQLiteRows.DeclTypes method is rarely called. It also changes the benchmark suite to use an in-memory database since we do not want the file system interfering with benchmark results. goos: darwin goarch: arm64 pkg: github.com/mattn/go-sqlite3 cpu: Apple M1 Max │ base.10.txt │ new.10.txt │ │ sec/op │ sec/op vs base │ CustomFunctions-10 3.318µ ± 2% 3.115µ ± 2% -6.10% (p=0.000 n=10) Suite/BenchmarkExec-10 1.236µ ± 1% 1.240µ ± 2% ~ (p=0.617 n=10) Suite/BenchmarkQuery-10 4.004µ ± 7% 3.363µ ± 2% -16.02% (p=0.000 n=10) Suite/BenchmarkParams-10 4.241µ ± 1% 3.758µ ± 2% -11.40% (p=0.000 n=10) Suite/BenchmarkStmt-10 2.830µ ± 0% 2.378µ ± 2% -15.97% (p=0.000 n=10) Suite/BenchmarkRows-10 126.3µ ± 1% 101.3µ ± 1% -19.79% (p=0.000 n=10) Suite/BenchmarkStmtRows-10 124.9µ ± 1% 100.5µ ± 2% -19.56% (p=0.000 n=10) Suite/BenchmarkStmt10Cols-10 10.130µ ± 0% 7.042µ ± 1% -30.48% (p=0.000 n=10) geomean 8.655µ 7.328µ -15.33% │ base.10.txt │ new.10.txt │ │ B/op │ B/op vs base │ CustomFunctions-10 568.0 ± 0% 576.0 ± 0% +1.41% (p=0.000 n=10) Suite/BenchmarkExec-10 128.0 ± 0% 128.0 ± 0% ~ (p=1.000 n=10) ¹ Suite/BenchmarkQuery-10 688.0 ± 0% 648.0 ± 0% -5.81% (p=0.000 n=10) Suite/BenchmarkParams-10 1.078Ki ± 0% 1.031Ki ± 0% -4.35% (p=0.000 n=10) Suite/BenchmarkStmt-10 920.0 ± 0% 872.0 ± 0% -5.22% (p=0.000 n=10) Suite/BenchmarkRows-10 9.305Ki ± 0% 9.188Ki ± 0% -1.26% (p=0.000 n=10) Suite/BenchmarkStmtRows-10 9.289Ki ± 0% 9.164Ki ± 0% -1.35% (p=0.000 n=10) Suite/BenchmarkStmt10Cols-10 992.0 ± 0% 696.0 ± 0% -29.84% (p=0.000 n=10) geomean 1.181Ki 1.106Ki -6.35% ¹ all samples are equal │ base.10.txt │ new.10.txt │ │ allocs/op │ allocs/op vs base │ CustomFunctions-10 18.00 ± 0% 18.00 ± 0% ~ (p=1.000 n=10) ¹ Suite/BenchmarkExec-10 7.000 ± 0% 7.000 ± 0% ~ (p=1.000 n=10) ¹ Suite/BenchmarkQuery-10 23.00 ± 0% 23.00 ± 0% ~ (p=1.000 n=10) ¹ Suite/BenchmarkParams-10 27.00 ± 0% 27.00 ± 0% ~ (p=1.000 n=10) ¹ Suite/BenchmarkStmt-10 25.00 ± 0% 25.00 ± 0% ~ (p=1.000 n=10) ¹ Suite/BenchmarkRows-10 525.0 ± 0% 519.0 ± 0% -1.14% (p=0.000 n=10) Suite/BenchmarkStmtRows-10 524.0 ± 0% 518.0 ± 0% -1.15% (p=0.000 n=10) Suite/BenchmarkStmt10Cols-10 39.00 ± 0% 19.00 ± 0% -51.28% (p=0.000 n=10) geomean 46.26 42.17 -8.86% ¹ all samples are equal
1 parent 82bc911 commit 4fb2899

File tree

2 files changed

+165
-29
lines changed

2 files changed

+165
-29
lines changed

‎sqlite3.go‎

Lines changed: 93 additions & 22 deletions
Original file line numberDiff line numberDiff line change
@@ -201,6 +201,53 @@ static int sqlite3_system_errno(sqlite3 *db) {
201201
return 0;
202202
}
203203
#endif
204+
205+
#define GO_SQLITE3_DECL_DATE (1 << 7)
206+
#define GO_SQLITE3_DECL_BOOL (1 << 6)
207+
#define GO_SQLITE3_DECL_MASK (GO_SQLITE3_DECL_DATE | GO_SQLITE3_DECL_BOOL)
208+
#define GO_SQLITE3_TYPE_MASK (GO_SQLITE3_DECL_BOOL - 1)
209+
210+
// _sqlite3_column_decltypes stores the declared column type in the typs array.
211+
// This function must always be called before _sqlite3_column_types since it
212+
// overwrites the datatype.
213+
static void _sqlite3_column_decltypes(sqlite3_stmt* stmt, uint8_t *typs, int ntyps) {
214+
for (int i = 0; i < ntyps; i++) {
215+
const char *typ = sqlite3_column_decltype(stmt, i);
216+
if (typ == NULL) {
217+
typs[i] = 0;
218+
continue;
219+
}
220+
switch (typ[0]) {
221+
case 'b':
222+
case 'B':
223+
if (!sqlite3_stricmp(typ, "boolean")) {
224+
typs[i] = GO_SQLITE3_DECL_BOOL;
225+
}
226+
break;
227+
case 'd':
228+
case 'D':
229+
if (!sqlite3_stricmp(typ, "date") || !sqlite3_stricmp(typ, "datetime")) {
230+
typs[i] = GO_SQLITE3_DECL_DATE;
231+
}
232+
break;
233+
case 't':
234+
case 'T':
235+
if (!sqlite3_stricmp(typ, "timestamp")) {
236+
typs[i] = GO_SQLITE3_DECL_DATE;
237+
}
238+
break;
239+
default:
240+
typs[i] = 0;
241+
}
242+
}
243+
}
244+
245+
static void _sqlite3_column_types(sqlite3_stmt *stmt, uint8_t *typs, int ntyps) {
246+
for (int i = 0; i < ntyps; i++) {
247+
typs[i] &= GO_SQLITE3_DECL_MASK; // clear lower bits
248+
typs[i] |= (uint8_t)sqlite3_column_type(stmt, i);
249+
}
250+
}
204251
*/
205252
import "C"
206253
import (
@@ -239,12 +286,6 @@ var SQLiteTimestampFormats = []string{
239286
"2006-01-02",
240287
}
241288

242-
const (
243-
columnDate string = "date"
244-
columnDatetime string = "datetime"
245-
columnTimestamp string = "timestamp"
246-
)
247-
248289
// This variable can be replaced with -ldflags like below:
249290
// go build -ldflags="-X 'github.com/mattn/go-sqlite3.driverName=my-sqlite3'"
250291
var driverName = "sqlite3"
@@ -390,12 +431,31 @@ type SQLiteResult struct {
390431
changes int64
391432
}
392433

434+
// A columnType is a compact representation of sqlite3 columns datatype and
435+
// declared type. The first two bits store the declared type and the remaining
436+
// six bits store the sqlite3 datatype.
437+
type columnType uint8
438+
439+
// DeclType returns the declared type, which is currently GO_SQLITE3_DECL_DATE
440+
// or GO_SQLITE3_DECL_BOOL, since those are the only two types that we need for
441+
// converting values.
442+
func (c columnType) DeclType() int {
443+
return int(c) & C.GO_SQLITE3_DECL_MASK
444+
}
445+
446+
// DataType returns the sqlite3 datatype code of the column, which is the
447+
// result of sqlite3_column_type.
448+
func (c columnType) DataType() int {
449+
return int(c) & C.GO_SQLITE3_TYPE_MASK
450+
}
451+
393452
// SQLiteRows implements driver.Rows.
394453
type SQLiteRows struct {
395454
s *SQLiteStmt
396455
nc int
397456
cols []string
398457
decltype []string
458+
coltype []columnType
399459
cls bool
400460
closed bool
401461
ctx context.Context // no better alternative to pass context into Next() method
@@ -2146,7 +2206,10 @@ func (rc *SQLiteRows) Columns() []string {
21462206
return rc.cols
21472207
}
21482208

2149-
func (rc *SQLiteRows) declTypes() []string {
2209+
// DeclTypes return column types.
2210+
func (rc *SQLiteRows) DeclTypes() []string {
2211+
rc.s.mu.Lock()
2212+
defer rc.s.mu.Unlock()
21502213
if rc.s.s != nil && rc.decltype == nil {
21512214
rc.decltype = make([]string, rc.nc)
21522215
for i := 0; i < rc.nc; i++ {
@@ -2156,13 +2219,6 @@ func (rc *SQLiteRows) declTypes() []string {
21562219
return rc.decltype
21572220
}
21582221

2159-
// DeclTypes return column types.
2160-
func (rc *SQLiteRows) DeclTypes() []string {
2161-
rc.s.mu.Lock()
2162-
defer rc.s.mu.Unlock()
2163-
return rc.declTypes()
2164-
}
2165-
21662222
// Next move cursor to next. Attempts to honor context timeout from QueryContext call.
21672223
func (rc *SQLiteRows) Next(dest []driver.Value) error {
21682224
rc.s.mu.Lock()
@@ -2195,6 +2251,13 @@ func (rc *SQLiteRows) Next(dest []driver.Value) error {
21952251
}
21962252
}
21972253

2254+
func (rc *SQLiteRows) colTypePtr() *C.uint8_t {
2255+
if len(rc.coltype) == 0 {
2256+
return nil
2257+
}
2258+
return (*C.uint8_t)(unsafe.Pointer(&rc.coltype[0]))
2259+
}
2260+
21982261
// nextSyncLocked moves cursor to next; must be called with locked mutex.
21992262
func (rc *SQLiteRows) nextSyncLocked(dest []driver.Value) error {
22002263
rv := C._sqlite3_step_internal(rc.s.s)
@@ -2208,15 +2271,24 @@ func (rc *SQLiteRows) nextSyncLocked(dest []driver.Value) error {
22082271
}
22092272
return nil
22102273
}
2274+
if len(dest) == 0 {
2275+
return nil
2276+
}
22112277

2212-
rc.declTypes()
2278+
if rc.coltype == nil {
2279+
rc.coltype = make([]columnType, rc.nc)
2280+
C._sqlite3_column_decltypes(rc.s.s, rc.colTypePtr(), C.int(rc.nc))
2281+
}
2282+
// Must call this each time since sqlite3 is loosely
2283+
// typed and the column types can vary between rows.
2284+
C._sqlite3_column_types(rc.s.s, rc.colTypePtr(), C.int(rc.nc))
22132285

22142286
for i := range dest {
2215-
switch C.sqlite3_column_type(rc.s.s, C.int(i)) {
2287+
switch rc.coltype[i].DataType() {
22162288
case C.SQLITE_INTEGER:
22172289
val := int64(C.sqlite3_column_int64(rc.s.s, C.int(i)))
2218-
switch rc.decltype[i] {
2219-
case columnTimestamp, columnDatetime, columnDate:
2290+
switch rc.coltype[i].DeclType() {
2291+
case C.GO_SQLITE3_DECL_DATE:
22202292
var t time.Time
22212293
// Assume a millisecond unix timestamp if it's 13 digits -- too
22222294
// large to be a reasonable timestamp in seconds.
@@ -2231,7 +2303,7 @@ func (rc *SQLiteRows) nextSyncLocked(dest []driver.Value) error {
22312303
t = t.In(rc.s.c.loc)
22322304
}
22332305
dest[i] = t
2234-
case "boolean":
2306+
case C.GO_SQLITE3_DECL_BOOL:
22352307
dest[i] = val > 0
22362308
default:
22372309
dest[i] = val
@@ -2255,8 +2327,7 @@ func (rc *SQLiteRows) nextSyncLocked(dest []driver.Value) error {
22552327
n := int(C.sqlite3_column_bytes(rc.s.s, C.int(i)))
22562328
s := C.GoStringN((*C.char)(unsafe.Pointer(C.sqlite3_column_text(rc.s.s, C.int(i)))), C.int(n))
22572329

2258-
switch rc.decltype[i] {
2259-
case columnTimestamp, columnDatetime, columnDate:
2330+
if rc.coltype[i].DeclType() == C.GO_SQLITE3_DECL_DATE {
22602331
var t time.Time
22612332
s = strings.TrimSuffix(s, "Z")
22622333
for _, format := range SQLiteTimestampFormats {
@@ -2273,7 +2344,7 @@ func (rc *SQLiteRows) nextSyncLocked(dest []driver.Value) error {
22732344
t = t.In(rc.s.c.loc)
22742345
}
22752346
dest[i] = t
2276-
default:
2347+
} else {
22772348
dest[i] = s
22782349
}
22792350
}

‎sqlite3_test.go‎

Lines changed: 72 additions & 7 deletions
Original file line numberDiff line numberDiff line change
@@ -2030,7 +2030,7 @@ func BenchmarkCustomFunctions(b *testing.B) {
20302030
}
20312031

20322032
func TestSuite(t *testing.T) {
2033-
initializeTestDB(t)
2033+
initializeTestDB(t, false)
20342034
defer freeTestDB()
20352035

20362036
for _, test := range tests {
@@ -2039,7 +2039,7 @@ func TestSuite(t *testing.T) {
20392039
}
20402040

20412041
func BenchmarkSuite(b *testing.B) {
2042-
initializeTestDB(b)
2042+
initializeTestDB(b, true)
20432043
defer freeTestDB()
20442044

20452045
for _, benchmark := range benchmarks {
@@ -2068,8 +2068,13 @@ type TestDB struct {
20682068

20692069
var db *TestDB
20702070

2071-
func initializeTestDB(t testing.TB) {
2072-
tempFilename := TempFilename(t)
2071+
func initializeTestDB(t testing.TB, memory bool) {
2072+
var tempFilename string
2073+
if memory {
2074+
tempFilename = ":memory:"
2075+
} else {
2076+
tempFilename = TempFilename(t)
2077+
}
20732078
d, err := sql.Open("sqlite3", tempFilename+"?_busy_timeout=99999")
20742079
if err != nil {
20752080
os.Remove(tempFilename)
@@ -2084,9 +2089,11 @@ func freeTestDB() {
20842089
if err != nil {
20852090
panic(err)
20862091
}
2087-
err = os.Remove(db.tempFilename)
2088-
if err != nil {
2089-
panic(err)
2092+
if db.tempFilename != "" && db.tempFilename != ":memory:" {
2093+
err := os.Remove(db.tempFilename)
2094+
if err != nil {
2095+
panic(err)
2096+
}
20902097
}
20912098
}
20922099

@@ -2111,6 +2118,7 @@ var benchmarks = []testing.InternalBenchmark{
21112118
{Name: "BenchmarkStmt", F: benchmarkStmt},
21122119
{Name: "BenchmarkRows", F: benchmarkRows},
21132120
{Name: "BenchmarkStmtRows", F: benchmarkStmtRows},
2121+
{Name: "BenchmarkStmt10Cols", F: benchmarkStmt10Cols},
21142122
}
21152123

21162124
func (db *TestDB) mustExec(sql string, args ...any) sql.Result {
@@ -2568,3 +2576,60 @@ func benchmarkStmtRows(b *testing.B) {
25682576
}
25692577
}
25702578
}
2579+
2580+
func benchmarkStmt10Cols(b *testing.B) {
2581+
db.once.Do(makeBench)
2582+
2583+
const createTableStmt = `
2584+
DROP TABLE IF EXISTS bench_cols;
2585+
VACUUM;
2586+
CREATE TABLE bench_cols (
2587+
r0 INTEGER NOT NULL,
2588+
r1 INTEGER NOT NULL,
2589+
r2 INTEGER NOT NULL,
2590+
r3 INTEGER NOT NULL,
2591+
r4 INTEGER NOT NULL,
2592+
r5 INTEGER NOT NULL,
2593+
r6 INTEGER NOT NULL,
2594+
r7 INTEGER NOT NULL,
2595+
r8 INTEGER NOT NULL,
2596+
r9 INTEGER NOT NULL
2597+
);`
2598+
if _, err := db.Exec(createTableStmt); err != nil {
2599+
b.Fatal(err)
2600+
}
2601+
for i := int64(0); i < 4; i++ {
2602+
_, err := db.Exec("INSERT INTO bench_cols VALUES (?, ?, ?, ?, ?, ?, ?, ?, ?, ?);",
2603+
i, i, i, i, i, i, i, i, i, i)
2604+
if err != nil {
2605+
b.Fatal(err)
2606+
}
2607+
}
2608+
2609+
stmt, err := db.Prepare("SELECT * FROM bench_cols;")
2610+
if err != nil {
2611+
b.Fatal(err)
2612+
}
2613+
defer stmt.Close()
2614+
2615+
b.ResetTimer()
2616+
var (
2617+
v0, v1, v2, v3, v4 int64
2618+
v5, v6, v7, v8, v9 int64
2619+
)
2620+
for i := 0; i < b.N; i++ {
2621+
rows, err := stmt.Query()
2622+
if err != nil {
2623+
b.Fatal(err)
2624+
}
2625+
for rows.Next() {
2626+
err := rows.Scan(&v0, &v1, &v2, &v3, &v4, &v5, &v6, &v7, &v8, &v9)
2627+
if err != nil {
2628+
b.Fatal(err)
2629+
}
2630+
}
2631+
if err := rows.Err(); err != nil {
2632+
b.Fatal(err)
2633+
}
2634+
}
2635+
}

0 commit comments

Comments
 (0)