Skip to content

Commit 85c8525

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 d3c66c9 commit 85c8525

File tree

2 files changed

+151
-22
lines changed

2 files changed

+151
-22
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
@@ -2147,7 +2207,10 @@ func (rc *SQLiteRows) Columns() []string {
21472207
return rc.cols
21482208
}
21492209

2150-
func (rc *SQLiteRows) declTypes() []string {
2210+
// DeclTypes return column types.
2211+
func (rc *SQLiteRows) DeclTypes() []string {
2212+
rc.s.mu.Lock()
2213+
defer rc.s.mu.Unlock()
21512214
if rc.s.s != nil && rc.decltype == nil {
21522215
rc.decltype = make([]string, rc.nc)
21532216
for i := 0; i < rc.nc; i++ {
@@ -2157,13 +2220,6 @@ func (rc *SQLiteRows) declTypes() []string {
21572220
return rc.decltype
21582221
}
21592222

2160-
// DeclTypes return column types.
2161-
func (rc *SQLiteRows) DeclTypes() []string {
2162-
rc.s.mu.Lock()
2163-
defer rc.s.mu.Unlock()
2164-
return rc.declTypes()
2165-
}
2166-
21672223
// Next move cursor to next. Attempts to honor context timeout from QueryContext call.
21682224
func (rc *SQLiteRows) Next(dest []driver.Value) error {
21692225
rc.s.mu.Lock()
@@ -2201,6 +2257,13 @@ func (rc *SQLiteRows) Next(dest []driver.Value) error {
22012257
}
22022258
}
22032259

2260+
func (rc *SQLiteRows) colTypePtr() *C.uint8_t {
2261+
if len(rc.coltype) == 0 {
2262+
return nil
2263+
}
2264+
return (*C.uint8_t)(unsafe.Pointer(&rc.coltype[0]))
2265+
}
2266+
22042267
// nextSyncLocked moves cursor to next; must be called with locked mutex.
22052268
func (rc *SQLiteRows) nextSyncLocked(dest []driver.Value) error {
22062269
rv := C._sqlite3_step_internal(rc.s.s)
@@ -2214,15 +2277,24 @@ func (rc *SQLiteRows) nextSyncLocked(dest []driver.Value) error {
22142277
}
22152278
return nil
22162279
}
2280+
if len(dest) == 0 {
2281+
return nil
2282+
}
22172283

2218-
rc.declTypes()
2284+
if rc.coltype == nil {
2285+
rc.coltype = make([]columnType, rc.nc)
2286+
C._sqlite3_column_decltypes(rc.s.s, rc.colTypePtr(), C.int(rc.nc))
2287+
}
2288+
// Must call this each time since sqlite3 is loosely
2289+
// typed and the column types can vary between rows.
2290+
C._sqlite3_column_types(rc.s.s, rc.colTypePtr(), C.int(rc.nc))
22192291

22202292
for i := range dest {
2221-
switch C.sqlite3_column_type(rc.s.s, C.int(i)) {
2293+
switch rc.coltype[i].DataType() {
22222294
case C.SQLITE_INTEGER:
22232295
val := int64(C.sqlite3_column_int64(rc.s.s, C.int(i)))
2224-
switch rc.decltype[i] {
2225-
case columnTimestamp, columnDatetime, columnDate:
2296+
switch rc.coltype[i].DeclType() {
2297+
case C.GO_SQLITE3_DECL_DATE:
22262298
var t time.Time
22272299
// Assume a millisecond unix timestamp if it's 13 digits -- too
22282300
// large to be a reasonable timestamp in seconds.
@@ -2237,7 +2309,7 @@ func (rc *SQLiteRows) nextSyncLocked(dest []driver.Value) error {
22372309
t = t.In(rc.s.c.loc)
22382310
}
22392311
dest[i] = t
2240-
case "boolean":
2312+
case C.GO_SQLITE3_DECL_BOOL:
22412313
dest[i] = val > 0
22422314
default:
22432315
dest[i] = val
@@ -2261,8 +2333,7 @@ func (rc *SQLiteRows) nextSyncLocked(dest []driver.Value) error {
22612333
n := int(C.sqlite3_column_bytes(rc.s.s, C.int(i)))
22622334
s := C.GoStringN((*C.char)(unsafe.Pointer(C.sqlite3_column_text(rc.s.s, C.int(i)))), C.int(n))
22632335

2264-
switch rc.decltype[i] {
2265-
case columnTimestamp, columnDatetime, columnDate:
2336+
if rc.coltype[i].DeclType() == C.GO_SQLITE3_DECL_DATE {
22662337
var t time.Time
22672338
s = strings.TrimSuffix(s, "Z")
22682339
for _, format := range SQLiteTimestampFormats {
@@ -2279,7 +2350,7 @@ func (rc *SQLiteRows) nextSyncLocked(dest []driver.Value) error {
22792350
t = t.In(rc.s.c.loc)
22802351
}
22812352
dest[i] = t
2282-
default:
2353+
} else {
22832354
dest[i] = s
22842355
}
22852356
}

‎sqlite3_test.go

Lines changed: 58 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -2120,6 +2120,7 @@ var benchmarks = []testing.InternalBenchmark{
21202120
{Name: "BenchmarkStmt", F: benchmarkStmt},
21212121
{Name: "BenchmarkRows", F: benchmarkRows},
21222122
{Name: "BenchmarkStmtRows", F: benchmarkStmtRows},
2123+
{Name: "BenchmarkStmt10Cols", F: benchmarkStmt10Cols},
21232124
}
21242125

21252126
func (db *TestDB) mustExec(sql string, args ...any) sql.Result {
@@ -2636,3 +2637,60 @@ func benchmarkStmtRows(b *testing.B) {
26362637
}
26372638
}
26382639
}
2640+
2641+
func benchmarkStmt10Cols(b *testing.B) {
2642+
db.once.Do(makeBench)
2643+
2644+
const createTableStmt = `
2645+
DROP TABLE IF EXISTS bench_cols;
2646+
VACUUM;
2647+
CREATE TABLE bench_cols (
2648+
r0 INTEGER NOT NULL,
2649+
r1 INTEGER NOT NULL,
2650+
r2 INTEGER NOT NULL,
2651+
r3 INTEGER NOT NULL,
2652+
r4 INTEGER NOT NULL,
2653+
r5 INTEGER NOT NULL,
2654+
r6 INTEGER NOT NULL,
2655+
r7 INTEGER NOT NULL,
2656+
r8 INTEGER NOT NULL,
2657+
r9 INTEGER NOT NULL
2658+
);`
2659+
if _, err := db.Exec(createTableStmt); err != nil {
2660+
b.Fatal(err)
2661+
}
2662+
for i := int64(0); i < 4; i++ {
2663+
_, err := db.Exec("INSERT INTO bench_cols VALUES (?, ?, ?, ?, ?, ?, ?, ?, ?, ?);",
2664+
i, i, i, i, i, i, i, i, i, i)
2665+
if err != nil {
2666+
b.Fatal(err)
2667+
}
2668+
}
2669+
2670+
stmt, err := db.Prepare("SELECT * FROM bench_cols;")
2671+
if err != nil {
2672+
b.Fatal(err)
2673+
}
2674+
defer stmt.Close()
2675+
2676+
b.ResetTimer()
2677+
var (
2678+
v0, v1, v2, v3, v4 int64
2679+
v5, v6, v7, v8, v9 int64
2680+
)
2681+
for i := 0; i < b.N; i++ {
2682+
rows, err := stmt.Query()
2683+
if err != nil {
2684+
b.Fatal(err)
2685+
}
2686+
for rows.Next() {
2687+
err := rows.Scan(&v0, &v1, &v2, &v3, &v4, &v5, &v6, &v7, &v8, &v9)
2688+
if err != nil {
2689+
b.Fatal(err)
2690+
}
2691+
}
2692+
if err := rows.Err(); err != nil {
2693+
b.Fatal(err)
2694+
}
2695+
}
2696+
}

0 commit comments

Comments
 (0)