-
Notifications
You must be signed in to change notification settings - Fork 8
/
tsvreader.go
662 lines (599 loc) · 13.4 KB
/
tsvreader.go
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
package tsvreader
import (
"bytes"
"fmt"
"io"
"math"
"strconv"
"time"
"unsafe"
)
// New returns new Reader that reads TSV data from r.
func New(r io.Reader) *Reader {
var tr Reader
tr.Reset(r)
return &tr
}
// Reader reads tab-separated data.
//
// Call New for creating new TSV reader.
// Call Next before reading the next row.
//
// It is expected that columns are separated by tabs while rows
// are separated by newlines.
type Reader struct {
r io.Reader
rb []byte
rErr error
rBuf [4 << 10]byte
col int
row int
rowBuf []byte
b []byte
scratch []byte
err error
needUnescape bool
}
// Reset resets the reader for reading from r.
func (tr *Reader) Reset(r io.Reader) {
tr.r = r
tr.rb = nil
tr.rErr = nil
tr.col = 0
tr.row = 0
tr.rowBuf = nil
tr.b = nil
tr.scratch = tr.scratch[:0]
tr.err = nil
tr.needUnescape = false
}
// Error returns the last error.
func (tr *Reader) Error() error {
if tr.err == io.EOF {
return nil
}
return tr.err
}
// ResetError resets the current error, so the reader could proceed further.
func (tr *Reader) ResetError() {
tr.err = nil
}
// HasCols returns true if the current row contains unread columns.
//
// An empty row doesn't contain columns.
//
// This function may be used if TSV stream contains rows with different
// number of colums.
func (tr *Reader) HasCols() bool {
return len(tr.rowBuf) > 0 && tr.b != nil
}
// Next advances to the next row.
//
// Returns true if the next row does exist.
//
// Next must be called after reading all the columns on the previous row.
// Check Error after Next returns false.
//
// HasCols may be used for reading rows with variable number of columns.
func (tr *Reader) Next() bool {
if tr.err != nil {
return false
}
if tr.HasCols() {
tr.err = fmt.Errorf("row #%d %q contains unread columns: %q", tr.row, tr.rowBuf, tr.b)
return false
}
tr.row++
tr.col = 0
tr.rowBuf = nil
for {
if len(tr.rb) == 0 {
// Read buffer is empty. Attempt to fill it.
if tr.rErr != nil {
tr.err = tr.rErr
if tr.err != io.EOF {
tr.err = fmt.Errorf("cannot read row #%d: %s", tr.row, tr.err)
} else if len(tr.scratch) > 0 {
tr.err = fmt.Errorf("cannot find newline at the end of row #%d; row: %q", tr.row, tr.scratch)
}
return false
}
n, err := tr.r.Read(tr.rBuf[:])
tr.rb = tr.rBuf[:n]
tr.needUnescape = (bytes.IndexByte(tr.rb, '\\') >= 0)
tr.rErr = err
}
// Search for the end of the current row.
n := bytes.IndexByte(tr.rb, '\n')
if n >= 0 {
// Fast path: the row has been found.
b := tr.rb[:n]
tr.rb = tr.rb[n+1:]
if len(tr.scratch) > 0 {
tr.scratch = append(tr.scratch, b...)
b = tr.scratch
tr.scratch = tr.scratch[:0]
}
tr.rowBuf = b
tr.b = tr.rowBuf
return true
}
// Slow path: cannot find the end of row.
// Append tr.rb to tr.scratch and repeat.
tr.scratch = append(tr.scratch, tr.rb...)
tr.rb = nil
}
}
// Int returns the next int column value from the current row.
func (tr *Reader) Int() int {
if tr.err != nil {
return 0
}
b, err := tr.nextCol()
if err != nil {
tr.setColError("cannot read `int`", err)
return 0
}
n, err := strconv.Atoi(b2s(b))
if err != nil {
tr.setColError("cannot parse `int`", err)
return 0
}
return n
}
// Uint returns the next uint column value from the current row.
func (tr *Reader) Uint() uint {
if tr.err != nil {
return 0
}
b, err := tr.nextCol()
if err != nil {
tr.setColError("cannot read `uint`", err)
return 0
}
s := b2s(b)
// Fast path - attempt to use Atoi
n, err := strconv.Atoi(s)
if err == nil && n >= 0 {
return uint(n)
}
// Slow path - use ParseUint
nu, err := strconv.ParseUint(s, 10, strconv.IntSize)
if err != nil {
tr.setColError("cannot parse `uint`", err)
return 0
}
return uint(nu)
}
// Int32 returns the next int32 column value from the current row.
func (tr *Reader) Int32() int32 {
if tr.err != nil {
return 0
}
b, err := tr.nextCol()
if err != nil {
tr.setColError("cannot read `int32`", err)
return 0
}
s := b2s(b)
// Fast path - attempt to use Atoi
n, err := strconv.Atoi(s)
if err == nil && n >= math.MinInt32 && n <= math.MaxInt32 {
return int32(n)
}
// Slow path - use ParseInt
n32, err := strconv.ParseInt(s, 10, 32)
if err != nil {
tr.setColError("cannot parse `int32`", err)
return 0
}
return int32(n32)
}
// Uint32 returns the next uint32 column value from the current row.
func (tr *Reader) Uint32() uint32 {
if tr.err != nil {
return 0
}
b, err := tr.nextCol()
if err != nil {
tr.setColError("cannot read `uint32`", err)
return 0
}
s := b2s(b)
// Fast path - attempt to use Atoi
n, err := strconv.Atoi(s)
if err == nil && n >= 0 && n <= math.MaxUint32 {
return uint32(n)
}
// Slow path - use ParseUint
n32, err := strconv.ParseUint(s, 10, 32)
if err != nil {
tr.setColError("cannot parse `uint32`", err)
return 0
}
return uint32(n32)
}
// Int16 returns the next int16 column value from the current row.
func (tr *Reader) Int16() int16 {
if tr.err != nil {
return 0
}
b, err := tr.nextCol()
if err != nil {
tr.setColError("cannot read `int16`", err)
return 0
}
n, err := strconv.Atoi(b2s(b))
if err != nil {
tr.setColError("cannot parse `int16`", err)
return 0
}
if n < math.MinInt16 || n > math.MaxInt16 {
tr.setColError("cannot parse `int16`", fmt.Errorf("out of range"))
return 0
}
return int16(n)
}
// Uint16 returns the next uint16 column value from the current row.
func (tr *Reader) Uint16() uint16 {
if tr.err != nil {
return 0
}
b, err := tr.nextCol()
if err != nil {
tr.setColError("cannot read `uint16`", err)
return 0
}
n, err := strconv.Atoi(b2s(b))
if err != nil {
tr.setColError("cannot parse `uint16`", err)
return 0
}
if n < 0 {
tr.setColError("cannot parse `uint16`", fmt.Errorf("invalid syntax"))
return 0
}
if n > math.MaxUint16 {
tr.setColError("cannot parse `uint16`", fmt.Errorf("out of range"))
return 0
}
return uint16(n)
}
// Int8 returns the next int8 column value from the current row.
func (tr *Reader) Int8() int8 {
if tr.err != nil {
return 0
}
b, err := tr.nextCol()
if err != nil {
tr.setColError("cannot read `int8`", err)
return 0
}
n, err := strconv.Atoi(b2s(b))
if err != nil {
tr.setColError("cannot parse `int8`", err)
return 0
}
if n < math.MinInt8 || n > math.MaxInt8 {
tr.setColError("cannot parse `int8`", fmt.Errorf("out of range"))
return 0
}
return int8(n)
}
// Uint8 returns the next uint8 column value from the current row.
func (tr *Reader) Uint8() uint8 {
if tr.err != nil {
return 0
}
b, err := tr.nextCol()
if err != nil {
tr.setColError("cannot read `uint8`", err)
return 0
}
n, err := strconv.Atoi(b2s(b))
if err != nil {
tr.setColError("cannot parse `uint8`", err)
return 0
}
if n < 0 {
tr.setColError("cannot parse `uint8`", fmt.Errorf("invalid syntax"))
return 0
}
if n > math.MaxUint8 {
tr.setColError("cannot parse `uint8`", fmt.Errorf("out of range"))
return 0
}
return uint8(n)
}
// Int64 returns the next int64 column value from the current row.
func (tr *Reader) Int64() int64 {
if tr.err != nil {
return 0
}
b, err := tr.nextCol()
if err != nil {
tr.setColError("cannot read `int64`", err)
return 0
}
s := b2s(b)
// Fast path - attempt to use Atoi
n, err := strconv.Atoi(s)
if err == nil && int64(n) >= math.MinInt64 && int64(n) <= math.MaxInt64 {
return int64(n)
}
// Slow path - use ParseInt
n64, err := strconv.ParseInt(s, 10, 64)
if err != nil {
tr.setColError("cannot parse `int64`", err)
return 0
}
return n64
}
// Uint64 returns the next uint64 column value from the current row.
func (tr *Reader) Uint64() uint64 {
if tr.err != nil {
return 0
}
b, err := tr.nextCol()
if err != nil {
tr.setColError("cannot read `uint64`", err)
return 0
}
s := b2s(b)
// Fast path - attempt to use Atoi
n, err := strconv.Atoi(s)
if err == nil && n >= 0 && uint64(n) <= math.MaxUint64 {
return uint64(n)
}
// Slow path - use ParseUint
n64, err := strconv.ParseUint(s, 10, 64)
if err != nil {
tr.setColError("cannot parse `uint64`", err)
return 0
}
return n64
}
// Float32 returns the next float32 column value from the current row.
func (tr *Reader) Float32() float32 {
if tr.err != nil {
return 0
}
b, err := tr.nextCol()
if err != nil {
tr.setColError("cannot read `float32`", err)
return 0
}
s := b2s(b)
f32, err := strconv.ParseFloat(s, 32)
if err != nil {
tr.setColError("cannot parse `float32`", err)
return 0
}
return float32(f32)
}
// Float64 returns the next float64 column value from the current row.
func (tr *Reader) Float64() float64 {
if tr.err != nil {
return 0
}
b, err := tr.nextCol()
if err != nil {
tr.setColError("cannot read `float64`", err)
return 0
}
s := b2s(b)
f64, err := strconv.ParseFloat(s, 64)
if err != nil {
tr.setColError("cannot parse `float64`", err)
return 0
}
return f64
}
// SkipCol skips the next column from the current row.
func (tr *Reader) SkipCol() {
if tr.err != nil {
return
}
_, err := tr.nextCol()
if err != nil {
tr.setColError("cannot skip column", err)
}
}
// Bytes returns the next bytes column value from the current row.
//
// The returned value is valid until the next call to Reader.
func (tr *Reader) Bytes() []byte {
if tr.err != nil {
return nil
}
b, err := tr.nextCol()
if err != nil {
tr.setColError("cannot read `bytes`", err)
return nil
}
if !tr.needUnescape {
// Fast path - nothing to unescape.
return b
}
// Unescape b
n := bytes.IndexByte(b, '\\')
if n < 0 {
// Nothing to unescape in the current column.
return b
}
// Slow path - in-place unescaping compatible with ClickHouse.
n++
d := b[:n]
b = b[n:]
for len(b) > 0 {
switch b[0] {
case 'b':
d[len(d)-1] = '\b'
case 'f':
d[len(d)-1] = '\f'
case 'r':
d[len(d)-1] = '\r'
case 'n':
d[len(d)-1] = '\n'
case 't':
d[len(d)-1] = '\t'
case '0':
d[len(d)-1] = 0
case '\'':
d[len(d)-1] = '\''
case '\\':
d[len(d)-1] = '\\'
default:
d[len(d)-1] = b[0]
}
b = b[1:]
n = bytes.IndexByte(b, '\\')
if n < 0 {
d = append(d, b...)
break
}
n++
d = append(d, b[:n]...)
b = b[n:]
}
return d
}
// String returns the next string column value from the current row.
//
// String allocates memory. Use Bytes to avoid memory allocations.
func (tr *Reader) String() string {
return string(tr.Bytes())
}
// Date returns the next date column value from the current row.
//
// date must be in the format YYYY-MM-DD
func (tr *Reader) Date() time.Time {
if tr.err != nil {
return zeroTime
}
b, err := tr.nextCol()
if err != nil {
tr.setColError("cannot read `date`", err)
return zeroTime
}
s := b2s(b)
y, m, d, err := parseDate(s)
if err != nil {
tr.setColError("cannot parse `date`", err)
return zeroTime
}
if y == 0 && m == 0 && d == 0 {
// special case for ClickHouse
return zeroTime
}
return time.Date(y, time.Month(m), d, 0, 0, 0, 0, time.UTC)
}
// DateTime returns the next datetime column value from the current row.
//
// datetime must be in the format YYYY-MM-DD hh:mm:ss.
func (tr *Reader) DateTime() time.Time {
if tr.err != nil {
return zeroTime
}
b, err := tr.nextCol()
if err != nil {
tr.setColError("cannot read `datetime`", err)
return zeroTime
}
s := b2s(b)
dt, err := parseDateTime(s)
if err != nil {
tr.setColError("cannot parse `datetime`", err)
return zeroTime
}
return dt
}
func parseDateTime(s string) (time.Time, error) {
if len(s) != len("YYYY-MM-DD hh:mm:ss") {
return zeroTime, fmt.Errorf("too short datetime")
}
y, m, d, err := parseDate(s[:len("YYYY-MM-DD")])
if err != nil {
return zeroTime, err
}
s = s[len("YYYY-MM-DD"):]
if s[0] != ' ' || s[3] != ':' || s[6] != ':' {
return zeroTime, fmt.Errorf("invalid time format. Must be hh:mm:ss")
}
hS := s[1:3]
minS := s[4:6]
secS := s[7:]
h, err := strconv.Atoi(hS)
if err != nil {
return zeroTime, fmt.Errorf("invalid hour: %s", err)
}
min, err := strconv.Atoi(minS)
if err != nil {
return zeroTime, fmt.Errorf("invalid minute: %s", err)
}
sec, err := strconv.Atoi(secS)
if err != nil {
return zeroTime, fmt.Errorf("invalid second: %s", err)
}
if y == 0 && m == 0 && d == 0 {
// Special case for ClickHouse
return zeroTime, nil
}
return time.Date(y, time.Month(m), d, h, min, sec, 0, time.UTC), nil
}
func parseDate(s string) (y, m, d int, err error) {
if len(s) != len("YYYY-MM-DD") {
err = fmt.Errorf("too short date")
return
}
s = s[:len("YYYY-MM-DD")]
if s[4] != '-' && s[7] != '-' {
err = fmt.Errorf("invalid date format. Must be YYYY-MM-DD")
return
}
yS := s[:4]
mS := s[5:7]
dS := s[8:]
y, err = strconv.Atoi(yS)
if err != nil {
err = fmt.Errorf("invalid year: %s", err)
return
}
m, err = strconv.Atoi(mS)
if err != nil {
err = fmt.Errorf("invalid month: %s", err)
return
}
d, err = strconv.Atoi(dS)
if err != nil {
err = fmt.Errorf("invalid day: %s", err)
return
}
return y, m, d, nil
}
var zeroTime time.Time
func (tr *Reader) nextCol() ([]byte, error) {
if tr.row == 0 {
return nil, fmt.Errorf("missing Next call")
}
tr.col++
if tr.b == nil {
return nil, fmt.Errorf("no more columns")
}
n := bytes.IndexByte(tr.b, '\t')
if n < 0 {
// last column
b := tr.b
tr.b = nil
return b, nil
}
b := tr.b[:n]
tr.b = tr.b[n+1:]
return b, nil
}
func (tr *Reader) setColError(msg string, err error) {
tr.err = fmt.Errorf("%s at row #%d, col #%d %q: %s", msg, tr.row, tr.col, tr.rowBuf, err)
}
func b2s(b []byte) string {
return *(*string)(unsafe.Pointer(&b))
}