1 /* $OpenBSD: join.c,v 1.16 2003/06/10 22:20:47 deraadt Exp $ */ 2 3 /*- 4 * Copyright (c) 1991, 1993, 1994 5 * The Regents of the University of California. All rights reserved. 6 * 7 * This code is derived from software contributed to Berkeley by 8 * Steve Hayman of the Computer Science Department, Indiana University, 9 * Michiro Hikida and David Goodenough. 10 * 11 * Redistribution and use in source and binary forms, with or without 12 * modification, are permitted provided that the following conditions 13 * are met: 14 * 1. Redistributions of source code must retain the above copyright 15 * notice, this list of conditions and the following disclaimer. 16 * 2. Redistributions in binary form must reproduce the above copyright 17 * notice, this list of conditions and the following disclaimer in the 18 * documentation and/or other materials provided with the distribution. 19 * 3. Neither the name of the University nor the names of its contributors 20 * may be used to endorse or promote products derived from this software 21 * without specific prior written permission. 22 * 23 * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND 24 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE 25 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE 26 * ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE 27 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL 28 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS 29 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) 30 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT 31 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY 32 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF 33 * SUCH DAMAGE. 34 */ 35 36 #ifndef lint 37 static char copyright[] = 38 "@(#) Copyright (c) 1991, 1993, 1994\n\ 39 The Regents of the University of California. All rights reserved.\n"; 40 #endif /* not lint */ 41 42 #ifndef lint 43 /*static char sccsid[] = "@(#)join.c 8.6 (Berkeley) 5/4/95"; */ 44 static char rcsid[] = "$OpenBSD: join.c,v 1.16 2003/06/10 22:20:47 deraadt Exp $"; 45 #endif /* not lint */ 46 47 #include <sys/param.h> 48 49 #include <ctype.h> 50 #include <err.h> 51 #include <errno.h> 52 #include <stdio.h> 53 #include <stdlib.h> 54 #include <string.h> 55 #include <unistd.h> 56 57 /* 58 * There's a structure per input file which encapsulates the state of the 59 * file. We repeatedly read lines from each file until we've read in all 60 * the consecutive lines from the file with a common join field. Then we 61 * compare the set of lines with an equivalent set from the other file. 62 */ 63 typedef struct { 64 char *line; /* line */ 65 u_long linealloc; /* line allocated count */ 66 char **fields; /* line field(s) */ 67 u_long fieldcnt; /* line field(s) count */ 68 u_long fieldalloc; /* line field(s) allocated count */ 69 u_long cfieldc; /* current field count */ 70 long fpos; /* fpos of start of field */ 71 } LINE; 72 73 typedef struct { 74 FILE *fp; /* file descriptor */ 75 u_long joinf; /* join field (-1, -2, -j) */ 76 int unpair; /* output unpairable lines (-a) */ 77 int number; /* 1 for file 1, 2 for file 2 */ 78 LINE *set; /* set of lines with same field */ 79 int pushbool; /* if pushback is set */ 80 u_long pushback; /* line on the stack */ 81 u_long setcnt; /* set count */ 82 u_long setalloc; /* set allocated count */ 83 u_long setusedc; /* sets used */ 84 } INPUT; 85 INPUT input1 = { NULL, 0, 0, 1, NULL, 0, 0, 0, }, 86 input2 = { NULL, 0, 0, 2, NULL, 0, 0, 0, }; 87 88 typedef struct { 89 u_long filenum; /* file number */ 90 u_long fieldno; /* field number */ 91 } OLIST; 92 OLIST *olist; /* output field list */ 93 u_long olistcnt; /* output field list count */ 94 u_long olistalloc; /* output field allocated count */ 95 96 int joinout = 1; /* show lines with matched join fields (-v) */ 97 int needsep; /* need separator character */ 98 int spans = 1; /* span multiple delimiters (-t) */ 99 char *empty; /* empty field replacement string (-e) */ 100 char *tabchar = " \t"; /* delimiter characters (-t) */ 101 102 int cmp(LINE *, u_long, LINE *, u_long); 103 void fieldarg(char *); 104 void joinlines(INPUT *, INPUT *); 105 void obsolete(char **); 106 void outfield(LINE *, u_long, int); 107 void outoneline(INPUT *, LINE *); 108 void outtwoline(INPUT *, LINE *, INPUT *, LINE *); 109 void slurp(INPUT *); 110 void slurpit(INPUT *); 111 void usage(void); 112 113 int 114 main(int argc, char *argv[]) 115 { 116 INPUT *F1, *F2; 117 int aflag, ch, cval, vflag; 118 char *end; 119 120 F1 = &input1; 121 F2 = &input2; 122 123 aflag = vflag = 0; 124 obsolete(argv); 125 while ((ch = getopt(argc, argv, "\01a:e:j:1:2:o:t:v:")) != -1) { 126 switch (ch) { 127 case '\01': /* See comment in obsolete(). */ 128 aflag = 1; 129 F1->unpair = F2->unpair = 1; 130 break; 131 case '1': 132 if ((F1->joinf = strtol(optarg, &end, 10)) < 1) 133 errx(1, "-1 option field number less than 1"); 134 if (*end) 135 errx(1, "illegal field number -- %s", optarg); 136 --F1->joinf; 137 break; 138 case '2': 139 if ((F2->joinf = strtol(optarg, &end, 10)) < 1) 140 errx(1, "-2 option field number less than 1"); 141 if (*end) 142 errx(1, "illegal field number -- %s", optarg); 143 --F2->joinf; 144 break; 145 case 'a': 146 aflag = 1; 147 switch(strtol(optarg, &end, 10)) { 148 case 1: 149 F1->unpair = 1; 150 break; 151 case 2: 152 F2->unpair = 1; 153 break; 154 default: 155 errx(1, "-a option file number not 1 or 2"); 156 break; 157 } 158 if (*end) 159 errx(1, "illegal file number -- %s", optarg); 160 break; 161 case 'e': 162 empty = optarg; 163 break; 164 case 'j': 165 if ((F1->joinf = F2->joinf = strtol(optarg, &end, 10)) < 1) 166 errx(1, "-j option field number less than 1"); 167 if (*end) 168 errx(1, "illegal field number -- %s", optarg); 169 --F1->joinf; 170 --F2->joinf; 171 break; 172 case 'o': 173 fieldarg(optarg); 174 break; 175 case 't': 176 spans = 0; 177 if (strlen(tabchar = optarg) != 1) 178 errx(1, "illegal tab character specification"); 179 break; 180 case 'v': 181 vflag = 1; 182 joinout = 0; 183 switch (strtol(optarg, &end, 10)) { 184 case 1: 185 F1->unpair = 1; 186 break; 187 case 2: 188 F2->unpair = 1; 189 break; 190 default: 191 errx(1, "-v option file number not 1 or 2"); 192 break; 193 } 194 if (*end) 195 errx(1, "illegal file number -- %s", optarg); 196 break; 197 case '?': 198 default: 199 usage(); 200 } 201 } 202 argc -= optind; 203 argv += optind; 204 205 if (aflag && vflag) 206 errx(1, "the -a and -v options are mutually exclusive"); 207 208 if (argc != 2) 209 usage(); 210 211 /* Open the files; "-" means stdin. */ 212 if (!strcmp(*argv, "-")) 213 F1->fp = stdin; 214 else if ((F1->fp = fopen(*argv, "r")) == NULL) 215 err(1, "%s", *argv); 216 ++argv; 217 if (!strcmp(*argv, "-")) 218 F2->fp = stdin; 219 else if ((F2->fp = fopen(*argv, "r")) == NULL) 220 err(1, "%s", *argv); 221 if (F1->fp == stdin && F2->fp == stdin) 222 errx(1, "only one input file may be stdin"); 223 224 F1->setusedc = 0; 225 F2->setusedc = 0; 226 slurp(F1); 227 slurp(F2); 228 F1->set->cfieldc = 0; 229 F2->set->cfieldc = 0; 230 231 /* 232 * We try to let the files have the same field value, advancing 233 * whoever falls behind and always advancing the file(s) we output 234 * from. 235 */ 236 while (F1->setcnt && F2->setcnt) { 237 cval = cmp(F1->set, F1->joinf, F2->set, F2->joinf); 238 if (cval == 0) { 239 /* Oh joy, oh rapture, oh beauty divine! */ 240 if (joinout) 241 joinlines(F1, F2); 242 slurp(F1); 243 slurp(F2); 244 } 245 else { 246 if (F1->unpair 247 && (cval < 0 || F2->set->cfieldc == F2->setusedc -1)) { 248 joinlines(F1, NULL); 249 slurp(F1); 250 } 251 else if (cval < 0) 252 /* File 1 takes the lead... */ 253 slurp(F1); 254 if (F2->unpair 255 && (cval > 0 || F1->set->cfieldc == F1->setusedc -1)) { 256 joinlines(F2, NULL); 257 slurp(F2); 258 } 259 else if (cval > 0) 260 /* File 2 takes the lead... */ 261 slurp(F2); 262 } 263 } 264 265 /* 266 * Now that one of the files is used up, optionally output any 267 * remaining lines from the other file. 268 */ 269 if (F1->unpair) 270 while (F1->setcnt) { 271 joinlines(F1, NULL); 272 slurp(F1); 273 } 274 if (F2->unpair) 275 while (F2->setcnt) { 276 joinlines(F2, NULL); 277 slurp(F2); 278 } 279 280 return 0; 281 } 282 283 /* wrapper around slurpit() to keep track of what field we are on */ 284 void slurp(INPUT *F) 285 { 286 long fpos; 287 u_long cfieldc; 288 289 if (F->set == NULL) { 290 fpos = 0; 291 cfieldc = 0; 292 } 293 else { 294 fpos = F->set->fpos; 295 cfieldc = F->set->cfieldc; 296 } 297 slurpit(F); 298 if (F->set == NULL) 299 return; 300 else if (fpos != F->set->fpos) 301 F->set->cfieldc = cfieldc+1; 302 } 303 304 void 305 slurpit(INPUT *F) 306 { 307 LINE *lp, *lastlp, tmp; 308 size_t len; 309 int cnt; 310 char *bp, *fieldp; 311 long fpos; 312 /* 313 * Read all of the lines from an input file that have the same 314 * join field. 315 */ 316 317 F->setcnt = 0; 318 for (lastlp = NULL; ; ++F->setcnt, lastlp = lp) { 319 /* 320 * If we're out of space to hold line structures, allocate 321 * more. Initialize the structure so that we know that this 322 * is new space. 323 */ 324 if (F->setcnt == F->setalloc) { 325 cnt = F->setalloc; 326 F->setalloc += 50; 327 if ((F->set = realloc(F->set, 328 F->setalloc * sizeof(LINE))) == NULL) 329 err(1, NULL); 330 memset(F->set + cnt, 0, 50 * sizeof(LINE)); 331 /* re-set lastlp in case it moved */ 332 if (lastlp != NULL) 333 lastlp = &F->set[F->setcnt - 1]; 334 } 335 /* 336 * Get any pushed back line, else get the next line. Allocate 337 * space as necessary. If taking the line from the stack swap 338 * the two structures so that we don't lose space allocated to 339 * either structure. This could be avoided by doing another 340 * level of indirection, but it's probably okay as is. 341 * but it's probably okay as is. 342 */ 343 lp = &F->set[F->setcnt]; 344 if (F->pushbool) { 345 tmp = F->set[F->setcnt]; 346 F->set[F->setcnt] = F->set[F->pushback]; 347 F->set[F->pushback] = tmp; 348 F->pushbool = 0; 349 continue; 350 } 351 if ((bp = fgetln(F->fp, &len)) == NULL) 352 return; 353 /* 354 * we depend on knowing on what field we are, one safe way is 355 * the file position. 356 */ 357 fpos = ftell(F->fp) - len; 358 if (lp->linealloc <= len + 1) { 359 lp->linealloc += MAX(100, len + 1 - lp->linealloc); 360 if ((lp->line = realloc(lp->line, lp->linealloc)) 361 == NULL) 362 err(1, NULL); 363 } 364 F->setusedc++; 365 memmove(lp->line, bp, len); 366 lp->fpos = fpos; 367 /* Replace trailing newline, if it exists. */ 368 if (bp[len - 1] == '\n') 369 lp->line[len - 1] = '\0'; 370 else 371 lp->line[len] = '\0'; 372 bp = lp->line; 373 374 /* Split the line into fields, allocate space as necessary. */ 375 lp->fieldcnt = 0; 376 while ((fieldp = strsep(&bp, tabchar)) != NULL) { 377 if (spans && *fieldp == '\0') 378 continue; 379 if (lp->fieldcnt == lp->fieldalloc) { 380 lp->fieldalloc += 50; 381 if ((lp->fields = realloc(lp->fields, 382 lp->fieldalloc * sizeof(char *))) == NULL) 383 err(1, NULL); 384 } 385 lp->fields[lp->fieldcnt++] = fieldp; 386 } 387 388 /* See if the join field value has changed. */ 389 if (lastlp != NULL && cmp(lp, F->joinf, lastlp, F->joinf)) { 390 F->pushbool = 1; 391 F->pushback = F->setcnt; 392 break; 393 } 394 } 395 } 396 397 int 398 cmp(LINE *lp1, u_long fieldno1, LINE *lp2, u_long fieldno2) 399 { 400 if (lp1->fieldcnt <= fieldno1) 401 return (-1); 402 else if (lp2->fieldcnt <= fieldno2) 403 return (1); 404 return (strcmp(lp1->fields[fieldno1], lp2->fields[fieldno2])); 405 } 406 407 void 408 joinlines(INPUT *F1, INPUT *F2) 409 { 410 int cnt1, cnt2; 411 412 /* 413 * Output the results of a join comparison. The output may be from 414 * either file 1 or file 2 (in which case the first argument is the 415 * file from which to output) or from both. 416 */ 417 if (F2 == NULL) { 418 for (cnt1 = 0; cnt1 < F1->setcnt; ++cnt1) 419 outoneline(F1, &F1->set[cnt1]); 420 return; 421 } 422 for (cnt1 = 0; cnt1 < F1->setcnt; ++cnt1) 423 for (cnt2 = 0; cnt2 < F2->setcnt; ++cnt2) 424 outtwoline(F1, &F1->set[cnt1], F2, &F2->set[cnt2]); 425 } 426 427 void 428 outoneline(INPUT *F, LINE *lp) 429 { 430 int cnt; 431 432 /* 433 * Output a single line from one of the files, according to the 434 * join rules. This happens when we are writing unmatched single 435 * lines. Output empty fields in the right places. 436 */ 437 if (olist) 438 for (cnt = 0; cnt < olistcnt; ++cnt) { 439 if (olist[cnt].filenum == F->number) 440 outfield(lp, olist[cnt].fieldno, 0); 441 else 442 outfield(lp, 0, 1); 443 } 444 else 445 for (cnt = 0; cnt < lp->fieldcnt; ++cnt) 446 outfield(lp, cnt, 0); 447 putchar('\n'); 448 if (ferror(stdout)) 449 err(1, "stdout"); 450 needsep = 0; 451 } 452 453 void 454 outtwoline(INPUT *F1, LINE *lp1, INPUT *F2, LINE *lp2) 455 { 456 int cnt; 457 458 /* Output a pair of lines according to the join list (if any). */ 459 if (olist) { 460 for (cnt = 0; cnt < olistcnt; ++cnt) 461 if (olist[cnt].filenum == 1) 462 outfield(lp1, olist[cnt].fieldno, 0); 463 else /* if (olist[cnt].filenum == 2) */ 464 outfield(lp2, olist[cnt].fieldno, 0); 465 } else { 466 /* 467 * Output the join field, then the remaining fields from F1 468 * and F2. 469 */ 470 outfield(lp1, F1->joinf, 0); 471 for (cnt = 0; cnt < lp1->fieldcnt; ++cnt) 472 if (F1->joinf != cnt) 473 outfield(lp1, cnt, 0); 474 for (cnt = 0; cnt < lp2->fieldcnt; ++cnt) 475 if (F2->joinf != cnt) 476 outfield(lp2, cnt, 0); 477 } 478 putchar('\n'); 479 if (ferror(stdout)) 480 err(1, "stdout"); 481 needsep = 0; 482 } 483 484 void 485 outfield(LINE *lp, u_long fieldno, int out_empty) 486 { 487 if (needsep++) 488 putchar((int)*tabchar); 489 if (!ferror(stdout)) { 490 if (lp->fieldcnt <= fieldno || out_empty) { 491 if (empty != NULL) 492 fputs(empty, stdout); 493 } else { 494 if (*lp->fields[fieldno] == '\0') 495 return; 496 fputs(lp->fields[fieldno], stdout); 497 } 498 } 499 if (ferror(stdout)) 500 err(1, "stdout"); 501 } 502 503 /* 504 * Convert an output list argument "2.1, 1.3, 2.4" into an array of output 505 * fields. 506 */ 507 void 508 fieldarg(char *option) 509 { 510 u_long fieldno; 511 char *end, *token; 512 513 while ((token = strsep(&option, ", \t")) != NULL) { 514 if (*token == '\0') 515 continue; 516 if ((token[0] != '1' && token[0] != '2') || token[1] != '.') 517 errx(1, "malformed -o option field"); 518 fieldno = strtol(token + 2, &end, 10); 519 if (*end) 520 errx(1, "malformed -o option field"); 521 if (fieldno == 0) 522 errx(1, "field numbers are 1 based"); 523 if (olistcnt == olistalloc) { 524 olistalloc += 50; 525 if ((olist = realloc(olist, 526 olistalloc * sizeof(OLIST))) == NULL) 527 err(1, NULL); 528 } 529 olist[olistcnt].filenum = token[0] - '0'; 530 olist[olistcnt].fieldno = fieldno - 1; 531 ++olistcnt; 532 } 533 } 534 535 void 536 obsolete(char **argv) 537 { 538 int len; 539 char **p, *ap, *t; 540 541 while ((ap = *++argv) != NULL) { 542 /* Return if "--". */ 543 if (ap[0] == '-' && ap[1] == '-') 544 return; 545 /* skip if not an option */ 546 if (ap[0] != '-') 547 continue; 548 switch (ap[1]) { 549 case 'a': 550 /* 551 * The original join allowed "-a", which meant the 552 * same as -a1 plus -a2. POSIX 1003.2, Draft 11.2 553 * only specifies this as "-a 1" and "a -2", so we 554 * have to use another option flag, one that is 555 * unlikely to ever be used or accidentally entered 556 * on the command line. (Well, we could reallocate 557 * the argv array, but that hardly seems worthwhile.) 558 */ 559 if (ap[2] == '\0') 560 ap[1] = '\01'; 561 break; 562 case 'j': 563 /* 564 * The original join allowed "-j[12] arg" and "-j arg". 565 * Convert the former to "-[12] arg". Don't convert 566 * the latter since getopt(3) can handle it. 567 */ 568 switch(ap[2]) { 569 case '1': 570 if (ap[3] != '\0') 571 goto jbad; 572 ap[1] = '1'; 573 ap[2] = '\0'; 574 break; 575 case '2': 576 if (ap[3] != '\0') 577 goto jbad; 578 ap[1] = '2'; 579 ap[2] = '\0'; 580 break; 581 case '\0': 582 break; 583 default: 584 jbad: errx(1, "illegal option -- %s", ap); 585 usage(); 586 } 587 break; 588 case 'o': 589 /* 590 * The original join allowed "-o arg arg". 591 * Convert to "-o arg -o arg". 592 */ 593 if (ap[2] != '\0') 594 break; 595 for (p = argv + 2; *p != NULL; ++p) { 596 if ((p[0][0] != '1' && p[0][0] != '2') || p[0][1] != '.') 597 break; 598 len = strlen(*p); 599 if (len - 2 != strspn(*p + 2, "0123456789")) 600 break; 601 if ((t = malloc(len + 3)) == NULL) 602 err(1, NULL); 603 t[0] = '-'; 604 t[1] = 'o'; 605 memmove(t + 2, *p, len + 1); 606 *p = t; 607 } 608 argv = p - 1; 609 break; 610 } 611 } 612 } 613 614 void 615 usage(void) 616 { 617 (void)fprintf(stderr, "%s%s\n", 618 "usage: join [-a fileno | -v fileno ] [-e string] [-1 field] ", 619 "[-2 field]\n [-o list] [-t char] file1 file2"); 620 exit(1); 621 } 622