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
663
664
665
666
667
668
669
670
671
672
673
674
675
676
677
678
679
680
681
682
683
684
685
686
687
688
689
690
691
692
693
694
695
696
697
698
699
700
701
702
703
704
705
706
707
708
709
710
711
712
713
714
715
716
717
718
719
720
721
722
723
724
725
726
727
728
729
730
731
732
733
734
735
736
737
738
739
740
741
742
743
744
745
746
747
748
749
750
751
752
753
754
755
756
757
758
759
760
761
762
763
764
765
766
767
768
769
770
771
772
773
774
775
776
777
778
779
780
781
782
783
784
785
786
787
788
789
790
791
792
793
794
795
796
797
798
799
800
801
802
803
804
805
806
807
808
809
810
811
812
813
814
815
816
817
818
819
820
821
822
823
824
825
826
827
828
829
830
831
832
833
834
835
836
837
838
839
840
841
842
843
844
845
846
847
848
849
850
851
852
853
854
855
856
857
858
859
860
861
862
863
864
865
866
867
868
869
870
871
872
873
874
875
876
877
878
879
880
881
882
883
884
885
886
887
888
889
890
891
892
893
894
895
896
897
898
899
900
901
902
903
904
905
906
907
908
909
910
911
912
913
914
915
916
917
918
919
920
921
922
923
924
925
926
927
928
929
930
931
932
933
934
935
936
937
938
939
940
941
942
943
944
945
946
947
948
949
950
951
952
953
954
955
956
957
958
959
960
961
962
963
964
|
/*
** 2016-05-28
**
** The author disclaims copyright to this source code. In place of
** a legal notice, here is a blessing:
**
** May you do good and not evil.
** May you find forgiveness for yourself and forgive others.
** May you share freely, never taking more than you give.
**
******************************************************************************
**
** This file contains the implementation of an SQLite virtual table for
** reading CSV files.
**
** Usage:
**
** .load ./csv
** CREATE VIRTUAL TABLE temp.csv USING csv(filename=FILENAME);
** SELECT * FROM csv;
**
** The columns are named "c1", "c2", "c3", ... by default. Or the
** application can define its own CREATE TABLE statement using the
** schema= parameter, like this:
**
** CREATE VIRTUAL TABLE temp.csv2 USING csv(
** filename = "../http.log",
** schema = "CREATE TABLE x(date,ipaddr,url,referrer,userAgent)"
** );
**
** Instead of specifying a file, the text of the CSV can be loaded using
** the data= parameter.
**
** If the columns=N parameter is supplied, then the CSV file is assumed to have
** N columns. If both the columns= and schema= parameters are omitted, then
** the number and names of the columns is determined by the first line of
** the CSV input.
**
** Some extra debugging features (used for testing virtual tables) are available
** if this module is compiled with -DSQLITE_TEST.
*/
#include <sqlite3ext.h>
SQLITE_EXTENSION_INIT1
#include <string.h>
#include <stdlib.h>
#include <assert.h>
#include <stdarg.h>
#include <ctype.h>
#include <stdio.h>
#ifndef SQLITE_OMIT_VIRTUALTABLE
/*
** A macro to hint to the compiler that a function should not be
** inlined.
*/
#if defined(__GNUC__)
# define CSV_NOINLINE __attribute__((noinline))
#elif defined(_MSC_VER) && _MSC_VER>=1310
# define CSV_NOINLINE __declspec(noinline)
#else
# define CSV_NOINLINE
#endif
/* Max size of the error message in a CsvReader */
#define CSV_MXERR 200
/* Size of the CsvReader input buffer */
#define CSV_INBUFSZ 1024
/* A context object used when read a CSV file. */
typedef struct CsvReader CsvReader;
struct CsvReader {
FILE *in; /* Read the CSV text from this input stream */
char *z; /* Accumulated text for a field */
int n; /* Number of bytes in z */
int nAlloc; /* Space allocated for z[] */
int nLine; /* Current line number */
int bNotFirst; /* True if prior text has been seen */
int cTerm; /* Character that terminated the most recent field */
size_t iIn; /* Next unread character in the input buffer */
size_t nIn; /* Number of characters in the input buffer */
char *zIn; /* The input buffer */
char zErr[CSV_MXERR]; /* Error message */
};
/* Initialize a CsvReader object */
static void csv_reader_init(CsvReader *p){
p->in = 0;
p->z = 0;
p->n = 0;
p->nAlloc = 0;
p->nLine = 0;
p->bNotFirst = 0;
p->nIn = 0;
p->zIn = 0;
p->zErr[0] = 0;
}
/* Close and reset a CsvReader object */
static void csv_reader_reset(CsvReader *p){
if( p->in ){
fclose(p->in);
sqlite3_free(p->zIn);
}
sqlite3_free(p->z);
csv_reader_init(p);
}
/* Report an error on a CsvReader */
static void csv_errmsg(CsvReader *p, const char *zFormat, ...){
va_list ap;
va_start(ap, zFormat);
sqlite3_vsnprintf(CSV_MXERR, p->zErr, zFormat, ap);
va_end(ap);
}
/* Open the file associated with a CsvReader
** Return the number of errors.
*/
static int csv_reader_open(
CsvReader *p, /* The reader to open */
const char *zFilename, /* Read from this filename */
const char *zData /* ... or use this data */
){
if( zFilename ){
p->zIn = sqlite3_malloc( CSV_INBUFSZ );
if( p->zIn==0 ){
csv_errmsg(p, "out of memory");
return 1;
}
p->in = fopen(zFilename, "rb");
if( p->in==0 ){
sqlite3_free(p->zIn);
csv_reader_reset(p);
csv_errmsg(p, "cannot open '%s' for reading", zFilename);
return 1;
}
}else{
assert( p->in==0 );
p->zIn = (char*)zData;
p->nIn = strlen(zData);
}
return 0;
}
/* The input buffer has overflowed. Refill the input buffer, then
** return the next character
*/
static CSV_NOINLINE int csv_getc_refill(CsvReader *p){
size_t got;
assert( p->iIn>=p->nIn ); /* Only called on an empty input buffer */
assert( p->in!=0 ); /* Only called if reading froma file */
got = fread(p->zIn, 1, CSV_INBUFSZ, p->in);
if( got==0 ) return EOF;
p->nIn = got;
p->iIn = 1;
return p->zIn[0];
}
/* Return the next character of input. Return EOF at end of input. */
static int csv_getc(CsvReader *p){
if( p->iIn >= p->nIn ){
if( p->in!=0 ) return csv_getc_refill(p);
return EOF;
}
return ((unsigned char*)p->zIn)[p->iIn++];
}
/* Increase the size of p->z and append character c to the end.
** Return 0 on success and non-zero if there is an OOM error */
static CSV_NOINLINE int csv_resize_and_append(CsvReader *p, char c){
char *zNew;
int nNew = p->nAlloc*2 + 100;
zNew = sqlite3_realloc64(p->z, nNew);
if( zNew ){
p->z = zNew;
p->nAlloc = nNew;
p->z[p->n++] = c;
return 0;
}else{
csv_errmsg(p, "out of memory");
return 1;
}
}
/* Append a single character to the CsvReader.z[] array.
** Return 0 on success and non-zero if there is an OOM error */
static int csv_append(CsvReader *p, char c){
if( p->n>=p->nAlloc-1 ) return csv_resize_and_append(p, c);
p->z[p->n++] = c;
return 0;
}
/* Read a single field of CSV text. Compatible with rfc4180 and extended
** with the option of having a separator other than ",".
**
** + Input comes from p->in.
** + Store results in p->z of length p->n. Space to hold p->z comes
** from sqlite3_malloc64().
** + Keep track of the line number in p->nLine.
** + Store the character that terminates the field in p->cTerm. Store
** EOF on end-of-file.
**
** Return 0 at EOF or on OOM. On EOF, the p->cTerm character will have
** been set to EOF.
*/
static char *csv_read_one_field(CsvReader *p){
int c;
p->n = 0;
c = csv_getc(p);
if( c==EOF ){
p->cTerm = EOF;
return 0;
}
if( c=='"' ){
int pc, ppc;
int startLine = p->nLine;
pc = ppc = 0;
while( 1 ){
c = csv_getc(p);
if( c<='"' || pc=='"' ){
if( c=='\n' ) p->nLine++;
if( c=='"' ){
if( pc=='"' ){
pc = 0;
continue;
}
}
if( (c==',' && pc=='"')
|| (c=='\n' && pc=='"')
|| (c=='\n' && pc=='\r' && ppc=='"')
|| (c==EOF && pc=='"')
){
do{ p->n--; }while( p->z[p->n]!='"' );
p->cTerm = (char)c;
break;
}
if( pc=='"' && c!='\r' ){
csv_errmsg(p, "line %d: unescaped %c character", p->nLine, '"');
break;
}
if( c==EOF ){
csv_errmsg(p, "line %d: unterminated %c-quoted field\n",
startLine, '"');
p->cTerm = (char)c;
break;
}
}
if( csv_append(p, (char)c) ) return 0;
ppc = pc;
pc = c;
}
}else{
/* If this is the first field being parsed and it begins with the
** UTF-8 BOM (0xEF BB BF) then skip the BOM */
if( (c&0xff)==0xef && p->bNotFirst==0 ){
csv_append(p, (char)c);
c = csv_getc(p);
if( (c&0xff)==0xbb ){
csv_append(p, (char)c);
c = csv_getc(p);
if( (c&0xff)==0xbf ){
p->bNotFirst = 1;
p->n = 0;
return csv_read_one_field(p);
}
}
}
while( c>',' || (c!=EOF && c!=',' && c!='\n') ){
if( csv_append(p, (char)c) ) return 0;
c = csv_getc(p);
}
if( c=='\n' ){
p->nLine++;
if( p->n>0 && p->z[p->n-1]=='\r' ) p->n--;
}
p->cTerm = (char)c;
}
assert( p->z==0 || p->n<p->nAlloc );
if( p->z ) p->z[p->n] = 0;
p->bNotFirst = 1;
return p->z;
}
/* Forward references to the various virtual table methods implemented
** in this file. */
static int csvtabCreate(sqlite3*, void*, int, const char*const*,
sqlite3_vtab**,char**);
static int csvtabConnect(sqlite3*, void*, int, const char*const*,
sqlite3_vtab**,char**);
static int csvtabBestIndex(sqlite3_vtab*,sqlite3_index_info*);
static int csvtabDisconnect(sqlite3_vtab*);
static int csvtabOpen(sqlite3_vtab*, sqlite3_vtab_cursor**);
static int csvtabClose(sqlite3_vtab_cursor*);
static int csvtabFilter(sqlite3_vtab_cursor*, int idxNum, const char *idxStr,
int argc, sqlite3_value **argv);
static int csvtabNext(sqlite3_vtab_cursor*);
static int csvtabEof(sqlite3_vtab_cursor*);
static int csvtabColumn(sqlite3_vtab_cursor*,sqlite3_context*,int);
static int csvtabRowid(sqlite3_vtab_cursor*,sqlite3_int64*);
/* An instance of the CSV virtual table */
typedef struct CsvTable {
sqlite3_vtab base; /* Base class. Must be first */
char *zFilename; /* Name of the CSV file */
char *zData; /* Raw CSV data in lieu of zFilename */
long iStart; /* Offset to start of data in zFilename */
int nCol; /* Number of columns in the CSV file */
unsigned int tstFlags; /* Bit values used for testing */
} CsvTable;
/* Allowed values for tstFlags */
#define CSVTEST_FIDX 0x0001 /* Pretend that constrained searchs cost less*/
/* A cursor for the CSV virtual table */
typedef struct CsvCursor {
sqlite3_vtab_cursor base; /* Base class. Must be first */
CsvReader rdr; /* The CsvReader object */
char **azVal; /* Value of the current row */
int *aLen; /* Length of each entry */
sqlite3_int64 iRowid; /* The current rowid. Negative for EOF */
} CsvCursor;
/* Transfer error message text from a reader into a CsvTable */
static void csv_xfer_error(CsvTable *pTab, CsvReader *pRdr){
sqlite3_free(pTab->base.zErrMsg);
pTab->base.zErrMsg = sqlite3_mprintf("%s", pRdr->zErr);
}
/*
** This method is the destructor fo a CsvTable object.
*/
static int csvtabDisconnect(sqlite3_vtab *pVtab){
CsvTable *p = (CsvTable*)pVtab;
sqlite3_free(p->zFilename);
sqlite3_free(p->zData);
sqlite3_free(p);
return SQLITE_OK;
}
/* Skip leading whitespace. Return a pointer to the first non-whitespace
** character, or to the zero terminator if the string has only whitespace */
static const char *csv_skip_whitespace(const char *z){
while( isspace((unsigned char)z[0]) ) z++;
return z;
}
/* Remove trailing whitespace from the end of string z[] */
static void csv_trim_whitespace(char *z){
size_t n = strlen(z);
while( n>0 && isspace((unsigned char)z[n]) ) n--;
z[n] = 0;
}
/* Dequote the string */
static void csv_dequote(char *z){
int j;
char cQuote = z[0];
size_t i, n;
if( cQuote!='\'' && cQuote!='"' ) return;
n = strlen(z);
if( n<2 || z[n-1]!=z[0] ) return;
for(i=1, j=0; i<n-1; i++){
if( z[i]==cQuote && z[i+1]==cQuote ) i++;
z[j++] = z[i];
}
z[j] = 0;
}
/* Check to see if the string is of the form: "TAG = VALUE" with optional
** whitespace before and around tokens. If it is, return a pointer to the
** first character of VALUE. If it is not, return NULL.
*/
static const char *csv_parameter(const char *zTag, int nTag, const char *z){
z = csv_skip_whitespace(z);
if( strncmp(zTag, z, nTag)!=0 ) return 0;
z = csv_skip_whitespace(z+nTag);
if( z[0]!='=' ) return 0;
return csv_skip_whitespace(z+1);
}
/* Decode a parameter that requires a dequoted string.
**
** Return 1 if the parameter is seen, or 0 if not. 1 is returned
** even if there is an error. If an error occurs, then an error message
** is left in p->zErr. If there are no errors, p->zErr[0]==0.
*/
static int csv_string_parameter(
CsvReader *p, /* Leave the error message here, if there is one */
const char *zParam, /* Parameter we are checking for */
const char *zArg, /* Raw text of the virtual table argment */
char **pzVal /* Write the dequoted string value here */
){
const char *zValue;
zValue = csv_parameter(zParam,(int)strlen(zParam),zArg);
if( zValue==0 ) return 0;
p->zErr[0] = 0;
if( *pzVal ){
csv_errmsg(p, "more than one '%s' parameter", zParam);
return 1;
}
*pzVal = sqlite3_mprintf("%s", zValue);
if( *pzVal==0 ){
csv_errmsg(p, "out of memory");
return 1;
}
csv_trim_whitespace(*pzVal);
csv_dequote(*pzVal);
return 1;
}
/* Return 0 if the argument is false and 1 if it is true. Return -1 if
** we cannot really tell.
*/
static int csv_boolean(const char *z){
if( sqlite3_stricmp("yes",z)==0
|| sqlite3_stricmp("on",z)==0
|| sqlite3_stricmp("true",z)==0
|| (z[0]=='1' && z[1]==0)
){
return 1;
}
if( sqlite3_stricmp("no",z)==0
|| sqlite3_stricmp("off",z)==0
|| sqlite3_stricmp("false",z)==0
|| (z[0]=='0' && z[1]==0)
){
return 0;
}
return -1;
}
/* Check to see if the string is of the form: "TAG = BOOLEAN" or just "TAG".
** If it is, set *pValue to be the value of the boolean ("true" if there is
** not "= BOOLEAN" component) and return non-zero. If the input string
** does not begin with TAG, return zero.
*/
static int csv_boolean_parameter(
const char *zTag, /* Tag we are looking for */
int nTag, /* Size of the tag in bytes */
const char *z, /* Input parameter */
int *pValue /* Write boolean value here */
){
int b;
z = csv_skip_whitespace(z);
if( strncmp(zTag, z, nTag)!=0 ) return 0;
z = csv_skip_whitespace(z + nTag);
if( z[0]==0 ){
*pValue = 1;
return 1;
}
if( z[0]!='=' ) return 0;
z = csv_skip_whitespace(z+1);
b = csv_boolean(z);
if( b>=0 ){
*pValue = b;
return 1;
}
return 0;
}
/*
** Parameters:
** filename=FILENAME Name of file containing CSV content
** data=TEXT Direct CSV content.
** schema=SCHEMA Alternative CSV schema.
** header=YES|NO First row of CSV defines the names of
** columns if "yes". Default "no".
** columns=N Assume the CSV file contains N columns.
**
** Only available if compiled with SQLITE_TEST:
**
** testflags=N Bitmask of test flags. Optional
**
** If schema= is omitted, then the columns are named "c0", "c1", "c2",
** and so forth. If columns=N is omitted, then the file is opened and
** the number of columns in the first row is counted to determine the
** column count. If header=YES, then the first row is skipped.
*/
static int csvtabConnect(
sqlite3 *db,
void *pAux,
int argc, const char *const*argv,
sqlite3_vtab **ppVtab,
char **pzErr
){
CsvTable *pNew = 0; /* The CsvTable object to construct */
int bHeader = -1; /* header= flags. -1 means not seen yet */
int rc = SQLITE_OK; /* Result code from this routine */
int i, j; /* Loop counters */
#ifdef SQLITE_TEST
int tstFlags = 0; /* Value for testflags=N parameter */
#endif
int b; /* Value of a boolean parameter */
int nCol = -99; /* Value of the columns= parameter */
CsvReader sRdr; /* A CSV file reader used to store an error
** message and/or to count the number of columns */
static const char *azParam[] = {
"filename", "data", "schema",
};
char *azPValue[3]; /* Parameter values */
# define CSV_FILENAME (azPValue[0])
# define CSV_DATA (azPValue[1])
# define CSV_SCHEMA (azPValue[2])
assert( sizeof(azPValue)==sizeof(azParam) );
memset(&sRdr, 0, sizeof(sRdr));
memset(azPValue, 0, sizeof(azPValue));
for(i=3; i<argc; i++){
const char *z = argv[i];
const char *zValue;
for(j=0; j<sizeof(azParam)/sizeof(azParam[0]); j++){
if( csv_string_parameter(&sRdr, azParam[j], z, &azPValue[j]) ) break;
}
if( j<sizeof(azParam)/sizeof(azParam[0]) ){
if( sRdr.zErr[0] ) goto csvtab_connect_error;
}else
if( csv_boolean_parameter("header",6,z,&b) ){
if( bHeader>=0 ){
csv_errmsg(&sRdr, "more than one 'header' parameter");
goto csvtab_connect_error;
}
bHeader = b;
}else
#ifdef SQLITE_TEST
if( (zValue = csv_parameter("testflags",9,z))!=0 ){
tstFlags = (unsigned int)atoi(zValue);
}else
#endif
if( (zValue = csv_parameter("columns",7,z))!=0 ){
if( nCol>0 ){
csv_errmsg(&sRdr, "more than one 'columns' parameter");
goto csvtab_connect_error;
}
nCol = atoi(zValue);
if( nCol<=0 ){
csv_errmsg(&sRdr, "column= value must be positive");
goto csvtab_connect_error;
}
}else
{
csv_errmsg(&sRdr, "bad parameter: '%s'", z);
goto csvtab_connect_error;
}
}
if( (CSV_FILENAME==0)==(CSV_DATA==0) ){
csv_errmsg(&sRdr, "must specify either filename= or data= but not both");
goto csvtab_connect_error;
}
if( (nCol<=0 || bHeader==1)
&& csv_reader_open(&sRdr, CSV_FILENAME, CSV_DATA)
){
goto csvtab_connect_error;
}
pNew = sqlite3_malloc( sizeof(*pNew) );
*ppVtab = (sqlite3_vtab*)pNew;
if( pNew==0 ) goto csvtab_connect_oom;
memset(pNew, 0, sizeof(*pNew));
if( CSV_SCHEMA==0 ){
sqlite3_str *pStr = sqlite3_str_new(0);
char *zSep = "";
int iCol = 0;
sqlite3_str_appendf(pStr, "CREATE TABLE x(");
if( nCol<0 && bHeader<1 ){
nCol = 0;
do{
csv_read_one_field(&sRdr);
nCol++;
}while( sRdr.cTerm==',' );
}
if( nCol>0 && bHeader<1 ){
for(iCol=0; iCol<nCol; iCol++){
sqlite3_str_appendf(pStr, "%sc%d TEXT", zSep, iCol);
zSep = ",";
}
}else{
do{
char *z = csv_read_one_field(&sRdr);
if( (nCol>0 && iCol<nCol) || (nCol<0 && bHeader) ){
sqlite3_str_appendf(pStr,"%s\"%w\" TEXT", zSep, z);
zSep = ",";
iCol++;
}
}while( sRdr.cTerm==',' );
if( nCol<0 ){
nCol = iCol;
}else{
while( iCol<nCol ){
sqlite3_str_appendf(pStr,"%sc%d TEXT", zSep, ++iCol);
zSep = ",";
}
}
}
pNew->nCol = nCol;
sqlite3_str_appendf(pStr, ")");
CSV_SCHEMA = sqlite3_str_finish(pStr);
if( CSV_SCHEMA==0 ) goto csvtab_connect_oom;
}else if( nCol<0 ){
do{
csv_read_one_field(&sRdr);
pNew->nCol++;
}while( sRdr.cTerm==',' );
}else{
pNew->nCol = nCol;
}
pNew->zFilename = CSV_FILENAME; CSV_FILENAME = 0;
pNew->zData = CSV_DATA; CSV_DATA = 0;
#ifdef SQLITE_TEST
pNew->tstFlags = tstFlags;
#endif
if( bHeader!=1 ){
pNew->iStart = 0;
}else if( pNew->zData ){
pNew->iStart = (int)sRdr.iIn;
}else{
pNew->iStart = (int)(ftell(sRdr.in) - sRdr.nIn + sRdr.iIn);
}
csv_reader_reset(&sRdr);
rc = sqlite3_declare_vtab(db, CSV_SCHEMA);
if( rc ){
csv_errmsg(&sRdr, "bad schema: '%s' - %s", CSV_SCHEMA, sqlite3_errmsg(db));
goto csvtab_connect_error;
}
for(i=0; i<sizeof(azPValue)/sizeof(azPValue[0]); i++){
sqlite3_free(azPValue[i]);
}
/* Rationale for DIRECTONLY:
** An attacker who controls a database schema could use this vtab
** to exfiltrate sensitive data from other files in the filesystem.
** And, recommended practice is to put all CSV virtual tables in the
** TEMP namespace, so they should still be usable from within TEMP
** views, so there shouldn't be a serious loss of functionality by
** prohibiting the use of this vtab from persistent triggers and views.
*/
sqlite3_vtab_config(db, SQLITE_VTAB_DIRECTONLY);
return SQLITE_OK;
csvtab_connect_oom:
rc = SQLITE_NOMEM;
csv_errmsg(&sRdr, "out of memory");
csvtab_connect_error:
if( pNew ) csvtabDisconnect(&pNew->base);
for(i=0; i<sizeof(azPValue)/sizeof(azPValue[0]); i++){
sqlite3_free(azPValue[i]);
}
if( sRdr.zErr[0] ){
sqlite3_free(*pzErr);
*pzErr = sqlite3_mprintf("%s", sRdr.zErr);
}
csv_reader_reset(&sRdr);
if( rc==SQLITE_OK ) rc = SQLITE_ERROR;
return rc;
}
/*
** Reset the current row content held by a CsvCursor.
*/
static void csvtabCursorRowReset(CsvCursor *pCur){
CsvTable *pTab = (CsvTable*)pCur->base.pVtab;
int i;
for(i=0; i<pTab->nCol; i++){
sqlite3_free(pCur->azVal[i]);
pCur->azVal[i] = 0;
pCur->aLen[i] = 0;
}
}
/*
** The xConnect and xCreate methods do the same thing, but they must be
** different so that the virtual table is not an eponymous virtual table.
*/
static int csvtabCreate(
sqlite3 *db,
void *pAux,
int argc, const char *const*argv,
sqlite3_vtab **ppVtab,
char **pzErr
){
return csvtabConnect(db, pAux, argc, argv, ppVtab, pzErr);
}
/*
** Destructor for a CsvCursor.
*/
static int csvtabClose(sqlite3_vtab_cursor *cur){
CsvCursor *pCur = (CsvCursor*)cur;
csvtabCursorRowReset(pCur);
csv_reader_reset(&pCur->rdr);
sqlite3_free(cur);
return SQLITE_OK;
}
/*
** Constructor for a new CsvTable cursor object.
*/
static int csvtabOpen(sqlite3_vtab *p, sqlite3_vtab_cursor **ppCursor){
CsvTable *pTab = (CsvTable*)p;
CsvCursor *pCur;
size_t nByte;
nByte = sizeof(*pCur) + (sizeof(char*)+sizeof(int))*pTab->nCol;
pCur = sqlite3_malloc64( nByte );
if( pCur==0 ) return SQLITE_NOMEM;
memset(pCur, 0, nByte);
pCur->azVal = (char**)&pCur[1];
pCur->aLen = (int*)&pCur->azVal[pTab->nCol];
*ppCursor = &pCur->base;
if( csv_reader_open(&pCur->rdr, pTab->zFilename, pTab->zData) ){
csv_xfer_error(pTab, &pCur->rdr);
return SQLITE_ERROR;
}
return SQLITE_OK;
}
/*
** Advance a CsvCursor to its next row of input.
** Set the EOF marker if we reach the end of input.
*/
static int csvtabNext(sqlite3_vtab_cursor *cur){
CsvCursor *pCur = (CsvCursor*)cur;
CsvTable *pTab = (CsvTable*)cur->pVtab;
int i = 0;
char *z;
do{
z = csv_read_one_field(&pCur->rdr);
if( z==0 ){
break;
}
if( i<pTab->nCol ){
if( pCur->aLen[i] < pCur->rdr.n+1 ){
char *zNew = sqlite3_realloc64(pCur->azVal[i], pCur->rdr.n+1);
if( zNew==0 ){
csv_errmsg(&pCur->rdr, "out of memory");
csv_xfer_error(pTab, &pCur->rdr);
break;
}
pCur->azVal[i] = zNew;
pCur->aLen[i] = pCur->rdr.n+1;
}
memcpy(pCur->azVal[i], z, pCur->rdr.n+1);
i++;
}
}while( pCur->rdr.cTerm==',' );
if( z==0 && i==0 ){
pCur->iRowid = -1;
}else{
pCur->iRowid++;
while( i<pTab->nCol ){
sqlite3_free(pCur->azVal[i]);
pCur->azVal[i] = 0;
pCur->aLen[i] = 0;
i++;
}
}
return SQLITE_OK;
}
/*
** Return values of columns for the row at which the CsvCursor
** is currently pointing.
*/
static int csvtabColumn(
sqlite3_vtab_cursor *cur, /* The cursor */
sqlite3_context *ctx, /* First argument to sqlite3_result_...() */
int i /* Which column to return */
){
CsvCursor *pCur = (CsvCursor*)cur;
CsvTable *pTab = (CsvTable*)cur->pVtab;
if( i>=0 && i<pTab->nCol && pCur->azVal[i]!=0 ){
sqlite3_result_text(ctx, pCur->azVal[i], -1, SQLITE_TRANSIENT);
}
return SQLITE_OK;
}
/*
** Return the rowid for the current row.
*/
static int csvtabRowid(sqlite3_vtab_cursor *cur, sqlite_int64 *pRowid){
CsvCursor *pCur = (CsvCursor*)cur;
*pRowid = pCur->iRowid;
return SQLITE_OK;
}
/*
** Return TRUE if the cursor has been moved off of the last
** row of output.
*/
static int csvtabEof(sqlite3_vtab_cursor *cur){
CsvCursor *pCur = (CsvCursor*)cur;
return pCur->iRowid<0;
}
/*
** Only a full table scan is supported. So xFilter simply rewinds to
** the beginning.
*/
static int csvtabFilter(
sqlite3_vtab_cursor *pVtabCursor,
int idxNum, const char *idxStr,
int argc, sqlite3_value **argv
){
CsvCursor *pCur = (CsvCursor*)pVtabCursor;
CsvTable *pTab = (CsvTable*)pVtabCursor->pVtab;
pCur->iRowid = 0;
/* Ensure the field buffer is always allocated. Otherwise, if the
** first field is zero bytes in size, this may be mistaken for an OOM
** error in csvtabNext(). */
if( csv_append(&pCur->rdr, 0) ) return SQLITE_NOMEM;
if( pCur->rdr.in==0 ){
assert( pCur->rdr.zIn==pTab->zData );
assert( pTab->iStart>=0 );
assert( (size_t)pTab->iStart<=pCur->rdr.nIn );
pCur->rdr.iIn = pTab->iStart;
}else{
fseek(pCur->rdr.in, pTab->iStart, SEEK_SET);
pCur->rdr.iIn = 0;
pCur->rdr.nIn = 0;
}
return csvtabNext(pVtabCursor);
}
/*
** Only a forward full table scan is supported. xBestIndex is mostly
** a no-op. If CSVTEST_FIDX is set, then the presence of equality
** constraints lowers the estimated cost, which is fiction, but is useful
** for testing certain kinds of virtual table behavior.
*/
static int csvtabBestIndex(
sqlite3_vtab *tab,
sqlite3_index_info *pIdxInfo
){
pIdxInfo->estimatedCost = 1000000;
#ifdef SQLITE_TEST
if( (((CsvTable*)tab)->tstFlags & CSVTEST_FIDX)!=0 ){
/* The usual (and sensible) case is to always do a full table scan.
** The code in this branch only runs when testflags=1. This code
** generates an artifical and unrealistic plan which is useful
** for testing virtual table logic but is not helpful to real applications.
**
** Any ==, LIKE, or GLOB constraint is marked as usable by the virtual
** table (even though it is not) and the cost of running the virtual table
** is reduced from 1 million to just 10. The constraints are *not* marked
** as omittable, however, so the query planner should still generate a
** plan that gives a correct answer, even if they plan is not optimal.
*/
int i;
int nConst = 0;
for(i=0; i<pIdxInfo->nConstraint; i++){
unsigned char op;
if( pIdxInfo->aConstraint[i].usable==0 ) continue;
op = pIdxInfo->aConstraint[i].op;
if( op==SQLITE_INDEX_CONSTRAINT_EQ
|| op==SQLITE_INDEX_CONSTRAINT_LIKE
|| op==SQLITE_INDEX_CONSTRAINT_GLOB
){
pIdxInfo->estimatedCost = 10;
pIdxInfo->aConstraintUsage[nConst].argvIndex = nConst+1;
nConst++;
}
}
}
#endif
return SQLITE_OK;
}
static sqlite3_module CsvModule = {
0, /* iVersion */
csvtabCreate, /* xCreate */
csvtabConnect, /* xConnect */
csvtabBestIndex, /* xBestIndex */
csvtabDisconnect, /* xDisconnect */
csvtabDisconnect, /* xDestroy */
csvtabOpen, /* xOpen - open a cursor */
csvtabClose, /* xClose - close a cursor */
csvtabFilter, /* xFilter - configure scan constraints */
csvtabNext, /* xNext - advance a cursor */
csvtabEof, /* xEof - check for end of scan */
csvtabColumn, /* xColumn - read data */
csvtabRowid, /* xRowid - read data */
0, /* xUpdate */
0, /* xBegin */
0, /* xSync */
0, /* xCommit */
0, /* xRollback */
0, /* xFindMethod */
0, /* xRename */
};
#ifdef SQLITE_TEST
/*
** For virtual table testing, make a version of the CSV virtual table
** available that has an xUpdate function. But the xUpdate always returns
** SQLITE_READONLY since the CSV file is not really writable.
*/
static int csvtabUpdate(sqlite3_vtab *p,int n,sqlite3_value**v,sqlite3_int64*x){
return SQLITE_READONLY;
}
static sqlite3_module CsvModuleFauxWrite = {
0, /* iVersion */
csvtabCreate, /* xCreate */
csvtabConnect, /* xConnect */
csvtabBestIndex, /* xBestIndex */
csvtabDisconnect, /* xDisconnect */
csvtabDisconnect, /* xDestroy */
csvtabOpen, /* xOpen - open a cursor */
csvtabClose, /* xClose - close a cursor */
csvtabFilter, /* xFilter - configure scan constraints */
csvtabNext, /* xNext - advance a cursor */
csvtabEof, /* xEof - check for end of scan */
csvtabColumn, /* xColumn - read data */
csvtabRowid, /* xRowid - read data */
csvtabUpdate, /* xUpdate */
0, /* xBegin */
0, /* xSync */
0, /* xCommit */
0, /* xRollback */
0, /* xFindMethod */
0, /* xRename */
};
#endif /* SQLITE_TEST */
#endif /* !defined(SQLITE_OMIT_VIRTUALTABLE) */
#ifdef _WIN32
__declspec(dllexport)
#endif
/*
** This routine is called when the extension is loaded. The new
** CSV virtual table module is registered with the calling database
** connection.
*/
int sqlite3_csv_init(
sqlite3 *db,
char **pzErrMsg,
const sqlite3_api_routines *pApi
){
#ifndef SQLITE_OMIT_VIRTUALTABLE
int rc;
SQLITE_EXTENSION_INIT2(pApi);
rc = sqlite3_create_module(db, "csv", &CsvModule, 0);
#ifdef SQLITE_TEST
if( rc==SQLITE_OK ){
rc = sqlite3_create_module(db, "csv_wr", &CsvModuleFauxWrite, 0);
}
#endif
return rc;
#else
return SQLITE_OK;
#endif
}
|