1 /* $NetBSD: join.c,v 1.29 2008/07/21 14:19:23 lukem Exp $ */ 2 3 /*- 4 * Copyright (c) 1991 The Regents of the University of California. 5 * All rights reserved. 6 * 7 * This code is derived from software contributed to Berkeley by 8 * Steve Hayman of Indiana University, Michiro Hikida and David 9 * 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 #if HAVE_NBTOOL_CONFIG_H 37 #include "nbtool_config.h" 38 #endif 39 40 #include <sys/cdefs.h> 41 #ifndef lint 42 __COPYRIGHT("@(#) Copyright (c) 1991\ 43 The Regents of the University of California. All rights reserved."); 44 #endif /* not lint */ 45 46 #ifndef lint 47 #if 0 48 static char sccsid[] = "from: @(#)join.c 5.1 (Berkeley) 11/18/91"; 49 #else 50 __RCSID("$NetBSD: join.c,v 1.29 2008/07/21 14:19:23 lukem Exp $"); 51 #endif 52 #endif /* not lint */ 53 54 #include <sys/types.h> 55 #include <ctype.h> 56 #include <err.h> 57 #include <errno.h> 58 #include <stdio.h> 59 #include <stdlib.h> 60 #include <string.h> 61 #include <unistd.h> 62 63 /* 64 * There's a structure per input file which encapsulates the state of the 65 * file. We repeatedly read lines from each file until we've read in all 66 * the consecutive lines from the file with a common join field. Then we 67 * compare the set of lines with an equivalent set from the other file. 68 */ 69 typedef struct { 70 char *line; /* line */ 71 u_long linealloc; /* line allocated count */ 72 char **fields; /* line field(s) */ 73 u_long fieldcnt; /* line field(s) count */ 74 u_long fieldalloc; /* line field(s) allocated count */ 75 } LINE; 76 77 LINE noline = {"", 0, 0, 0, 0}; /* arg for outfield if no line to output */ 78 79 typedef struct { 80 FILE *fp; /* file descriptor */ 81 u_long joinf; /* join field (-1, -2, -j) */ 82 int unpair; /* output unpairable lines (-a) */ 83 int number; /* 1 for file 1, 2 for file 2 */ 84 85 LINE *set; /* set of lines with same field */ 86 u_long pushback; /* line on the stack */ 87 u_long setcnt; /* set count */ 88 u_long setalloc; /* set allocated count */ 89 } INPUT; 90 INPUT input1 = { NULL, 0, 0, 1, NULL, -1, 0, 0, }, 91 input2 = { NULL, 0, 0, 2, NULL, -1, 0, 0, }; 92 93 typedef struct { 94 u_long fileno; /* file number */ 95 u_long fieldno; /* field number */ 96 } OLIST; 97 OLIST *olist; /* output field list */ 98 u_long olistcnt; /* output field list count */ 99 u_long olistalloc; /* output field allocated count */ 100 101 int joinout = 1; /* show lines with matched join fields (-v) */ 102 int needsep; /* need separator character */ 103 int spans = 1; /* span multiple delimiters (-t) */ 104 char *empty; /* empty field replacement string (-e) */ 105 char *tabchar = " \t"; /* delimiter characters (-t) */ 106 107 int cmp(LINE *, u_long, LINE *, u_long); 108 void enomem(void); 109 void fieldarg(char *); 110 void joinlines(INPUT *, INPUT *); 111 int main(int, char **); 112 void obsolete(char **); 113 void outfield(LINE *, u_long); 114 void outoneline(INPUT *, LINE *); 115 void outtwoline(INPUT *, LINE *, INPUT *, LINE *); 116 void slurp(INPUT *); 117 void usage(void); 118 119 int 120 main(int argc, char *argv[]) 121 { 122 INPUT *F1, *F2; 123 int aflag, ch, cval, vflag; 124 char *end; 125 126 F1 = &input1; 127 F2 = &input2; 128 129 aflag = vflag = 0; 130 obsolete(argv); 131 while ((ch = getopt(argc, argv, "\01a:e:j:1:2:o:t:v:")) != -1) { 132 switch (ch) { 133 case '\01': 134 aflag = 1; 135 F1->unpair = F2->unpair = 1; 136 break; 137 case '1': 138 if ((F1->joinf = strtol(optarg, &end, 10)) < 1) { 139 warnx("-1 option field number less than 1"); 140 usage(); 141 } 142 if (*end) { 143 warnx("illegal field number -- %s", optarg); 144 usage(); 145 } 146 --F1->joinf; 147 break; 148 case '2': 149 if ((F2->joinf = strtol(optarg, &end, 10)) < 1) { 150 warnx("-2 option field number less than 1"); 151 usage(); 152 } 153 if (*end) { 154 warnx("illegal field number -- %s", optarg); 155 usage(); 156 } 157 --F2->joinf; 158 break; 159 case 'a': 160 aflag = 1; 161 switch(strtol(optarg, &end, 10)) { 162 case 1: 163 F1->unpair = 1; 164 break; 165 case 2: 166 F2->unpair = 1; 167 break; 168 default: 169 warnx("-a option file number not 1 or 2"); 170 usage(); 171 break; 172 } 173 if (*end) { 174 warnx("illegal file number -- %s", optarg); 175 usage(); 176 } 177 break; 178 case 'e': 179 empty = optarg; 180 break; 181 case 'j': 182 if ((F1->joinf = F2->joinf = 183 strtol(optarg, &end, 10)) < 1) { 184 warnx("-j option field number less than 1"); 185 usage(); 186 } 187 if (*end) { 188 warnx("illegal field number -- %s", optarg); 189 usage(); 190 } 191 --F1->joinf; 192 --F2->joinf; 193 break; 194 case 'o': 195 fieldarg(optarg); 196 break; 197 case 't': 198 spans = 0; 199 if (strlen(tabchar = optarg) != 1) { 200 warnx("illegal tab character specification"); 201 usage(); 202 } 203 break; 204 case 'v': 205 vflag = 1; 206 joinout = 0; 207 switch(strtol(optarg, &end, 10)) { 208 case 1: 209 F1->unpair = 1; 210 break; 211 case 2: 212 F2->unpair = 1; 213 break; 214 default: 215 warnx("-v option file number not 1 or 2"); 216 usage(); 217 break; 218 } 219 if (*end) { 220 warnx("illegal file number -- %s", optarg); 221 usage(); 222 } 223 break; 224 case '?': 225 default: 226 usage(); 227 } 228 } 229 argc -= optind; 230 argv += optind; 231 232 if (aflag && vflag) 233 errx(1, "-a and -v options mutually exclusive"); 234 235 if (argc != 2) 236 usage(); 237 238 /* Open the files; "-" means stdin. */ 239 if (!strcmp(*argv, "-")) 240 F1->fp = stdin; 241 else if ((F1->fp = fopen(*argv, "r")) == NULL) 242 err(1, "%s", *argv); 243 ++argv; 244 if (!strcmp(*argv, "-")) 245 F2->fp = stdin; 246 else if ((F2->fp = fopen(*argv, "r")) == NULL) 247 err(1, "%s", *argv); 248 if (F1->fp == stdin && F2->fp == stdin) 249 errx(1, "only one input file may be stdin"); 250 251 slurp(F1); 252 slurp(F2); 253 while (F1->setcnt && F2->setcnt) { 254 cval = cmp(F1->set, F1->joinf, F2->set, F2->joinf); 255 if (cval == 0) { 256 /* Oh joy, oh rapture, oh beauty divine! */ 257 if (joinout) 258 joinlines(F1, F2); 259 slurp(F1); 260 slurp(F2); 261 } else if (cval < 0) { 262 /* File 1 takes the lead... */ 263 if (F1->unpair) 264 joinlines(F1, NULL); 265 slurp(F1); 266 } else { 267 /* File 2 takes the lead... */ 268 if (F2->unpair) 269 joinlines(F2, NULL); 270 slurp(F2); 271 } 272 } 273 274 /* 275 * Now that one of the files is used up, optionally output any 276 * remaining lines from the other file. 277 */ 278 if (F1->unpair) 279 while (F1->setcnt) { 280 joinlines(F1, NULL); 281 slurp(F1); 282 } 283 if (F1->fp != stdin) 284 fclose(F1->fp); 285 286 if (F2->unpair) 287 while (F2->setcnt) { 288 joinlines(F2, NULL); 289 slurp(F2); 290 } 291 if (F2->fp != stdin) 292 fclose(F2->fp); 293 294 return 0; 295 } 296 297 void 298 slurp(INPUT *F) 299 { 300 LINE *lp; 301 LINE tmp; 302 LINE *nline; 303 size_t len; 304 int cnt; 305 char *bp, *fieldp; 306 u_long nsize; 307 308 /* 309 * Read all of the lines from an input file that have the same 310 * join field. 311 */ 312 for (F->setcnt = 0;; ++F->setcnt) { 313 /* 314 * If we're out of space to hold line structures, allocate 315 * more. Initialize the structure so that we know that this 316 * is new space. 317 */ 318 if (F->setcnt == F->setalloc) { 319 cnt = F->setalloc; 320 if (F->setalloc == 0) 321 nsize = 64; 322 else 323 nsize = F->setalloc << 1; 324 if ((nline = realloc(F->set, 325 nsize * sizeof(LINE))) == NULL) 326 enomem(); 327 F->set = nline; 328 F->setalloc = nsize; 329 memset(F->set + cnt, 0, 330 (F->setalloc - cnt) * sizeof(LINE)); 331 } 332 333 /* 334 * Get any pushed back line, else get the next line. Allocate 335 * space as necessary. If taking the line from the stack swap 336 * the two structures so that we don't lose the allocated space. 337 * This could be avoided by doing another level of indirection, 338 * but it's probably okay as is. 339 */ 340 lp = &F->set[F->setcnt]; 341 if (F->pushback != -1) { 342 tmp = F->set[F->setcnt]; 343 F->set[F->setcnt] = F->set[F->pushback]; 344 F->set[F->pushback] = tmp; 345 F->pushback = -1; 346 continue; 347 } 348 if ((bp = fgetln(F->fp, &len)) == NULL) 349 return; 350 if (lp->linealloc <= len + 1) { 351 char *n; 352 353 if (lp->linealloc == 0) 354 nsize = 128; 355 else 356 nsize = lp->linealloc; 357 while (nsize <= len + 1) 358 nsize <<= 1; 359 if ((n = realloc(lp->line, 360 nsize * sizeof(char))) == NULL) 361 enomem(); 362 lp->line = n; 363 lp->linealloc = nsize; 364 } 365 memmove(lp->line, bp, len); 366 367 /* 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 char **n; 381 382 if (lp->fieldalloc == 0) 383 nsize = 16; 384 else 385 nsize = lp->fieldalloc << 1; 386 if ((n = realloc(lp->fields, 387 nsize * sizeof(char *))) == NULL) 388 enomem(); 389 lp->fields = n; 390 lp->fieldalloc = nsize; 391 } 392 lp->fields[lp->fieldcnt++] = fieldp; 393 } 394 395 /* See if the join field value has changed. */ 396 if (F->setcnt && cmp(lp, F->joinf, lp - 1, F->joinf)) { 397 F->pushback = F->setcnt; 398 break; 399 } 400 } 401 } 402 403 int 404 cmp(LINE *lp1, u_long fieldno1, LINE *lp2, u_long fieldno2) 405 { 406 407 if (lp1->fieldcnt <= fieldno1) 408 return (lp2->fieldcnt <= fieldno2 ? 0 : 1); 409 if (lp2->fieldcnt <= fieldno2) 410 return (-1); 411 return (strcmp(lp1->fields[fieldno1], lp2->fields[fieldno2])); 412 } 413 414 void 415 joinlines(INPUT *F1, INPUT *F2) 416 { 417 int cnt1, cnt2; 418 419 /* 420 * Output the results of a join comparison. The output may be from 421 * either file 1 or file 2 (in which case the first argument is the 422 * file from which to output) or from both. 423 */ 424 if (F2 == NULL) { 425 for (cnt1 = 0; cnt1 < F1->setcnt; ++cnt1) 426 outoneline(F1, &F1->set[cnt1]); 427 return; 428 } 429 for (cnt1 = 0; cnt1 < F1->setcnt; ++cnt1) 430 for (cnt2 = 0; cnt2 < F2->setcnt; ++cnt2) 431 outtwoline(F1, &F1->set[cnt1], F2, &F2->set[cnt2]); 432 } 433 434 void 435 outoneline(INPUT *F, LINE *lp) 436 { 437 int cnt; 438 439 /* 440 * Output a single line from one of the files, according to the 441 * join rules. This happens when we are writing unmatched single 442 * lines. Output empty fields in the right places. 443 */ 444 if (olist) 445 for (cnt = 0; cnt < olistcnt; ++cnt) { 446 if (olist[cnt].fileno == F->number) 447 outfield(lp, olist[cnt].fieldno); 448 else 449 outfield(&noline, 1); 450 } 451 else 452 for (cnt = 0; cnt < lp->fieldcnt; ++cnt) 453 outfield(lp, cnt); 454 (void)printf("\n"); 455 if (ferror(stdout)) 456 err(1, "stdout"); 457 needsep = 0; 458 } 459 460 void 461 outtwoline(INPUT *F1, LINE *lp1, INPUT *F2, LINE *lp2) 462 { 463 int cnt; 464 465 /* Output a pair of lines according to the join list (if any). */ 466 if (olist) { 467 for (cnt = 0; cnt < olistcnt; ++cnt) 468 if (olist[cnt].fileno == 1) 469 outfield(lp1, olist[cnt].fieldno); 470 else /* if (olist[cnt].fileno == 2) */ 471 outfield(lp2, olist[cnt].fieldno); 472 } else { 473 /* 474 * Output the join field, then the remaining fields from F1 475 * and F2. 476 */ 477 outfield(lp1, F1->joinf); 478 for (cnt = 0; cnt < lp1->fieldcnt; ++cnt) 479 if (F1->joinf != cnt) 480 outfield(lp1, cnt); 481 for (cnt = 0; cnt < lp2->fieldcnt; ++cnt) 482 if (F2->joinf != cnt) 483 outfield(lp2, cnt); 484 } 485 (void)printf("\n"); 486 if (ferror(stdout)) 487 err(1, "stdout"); 488 needsep = 0; 489 } 490 491 void 492 outfield(LINE *lp, u_long fieldno) 493 { 494 if (needsep++) 495 (void)printf("%c", *tabchar); 496 if (!ferror(stdout)) { 497 if (lp->fieldcnt <= fieldno) { 498 if (empty != NULL) 499 (void)printf("%s", empty); 500 } else { 501 if (*lp->fields[fieldno] == '\0') 502 return; 503 (void)printf("%s", lp->fields[fieldno]); 504 } 505 } 506 if (ferror(stdout)) 507 err(1, "stdout"); 508 } 509 510 /* 511 * Convert an output list argument "2.1, 1.3, 2.4" into an array of output 512 * fields. 513 */ 514 void 515 fieldarg(char *option) 516 { 517 u_long fieldno; 518 char *end, *token; 519 OLIST *n; 520 521 while ((token = strsep(&option, ", \t")) != NULL) { 522 if (*token == '\0') 523 continue; 524 if ((token[0] != '1' && token[0] != '2') || token[1] != '.') 525 errx(1, "malformed -o option field"); 526 fieldno = strtol(token + 2, &end, 10); 527 if (*end) 528 errx(1, "malformed -o option field"); 529 if (fieldno == 0) 530 errx(1, "field numbers are 1 based"); 531 if (olistcnt == olistalloc) { 532 if ((n = realloc(olist, 533 (olistalloc + 50) * sizeof(OLIST))) == NULL) 534 enomem(); 535 olist = n; 536 olistalloc += 50; 537 } 538 olist[olistcnt].fileno = token[0] - '0'; 539 olist[olistcnt].fieldno = fieldno - 1; 540 ++olistcnt; 541 } 542 } 543 544 void 545 obsolete(char **argv) 546 { 547 int len; 548 char **p, *ap, *t; 549 550 while ((ap = *++argv) != NULL) { 551 /* Return if "--". */ 552 if (ap[0] == '-' && ap[1] == '-') 553 return; 554 switch (ap[1]) { 555 case 'a': 556 /* 557 * The original join allowed "-a", which meant the 558 * same as -a1 plus -a2. POSIX 1003.2, Draft 11.2 559 * only specifies this as "-a 1" and "a -2", so we 560 * have to use another option flag, one that is 561 * unlikely to ever be used or accidentally entered 562 * on the command line. (Well, we could reallocate 563 * the argv array, but that hardly seems worthwhile.) 564 */ 565 if (ap[2] == '\0') 566 ap[1] = '\01'; 567 break; 568 case 'j': 569 /* 570 * The original join allowed "-j[12] arg" and "-j arg". 571 * Convert the former to "-[12] arg". Don't convert 572 * the latter since getopt(3) can handle it. 573 */ 574 switch(ap[2]) { 575 case '1': 576 if (ap[3] != '\0') 577 goto jbad; 578 ap[1] = '1'; 579 ap[2] = '\0'; 580 break; 581 case '2': 582 if (ap[3] != '\0') 583 goto jbad; 584 ap[1] = '2'; 585 ap[2] = '\0'; 586 break; 587 case '\0': 588 break; 589 default: 590 jbad: errx(1, "illegal option -- %s", ap); 591 usage(); 592 } 593 break; 594 case 'o': 595 /* 596 * The original join allowed "-o arg arg". Convert to 597 * "-o arg -o arg". 598 */ 599 if (ap[2] != '\0') 600 break; 601 for (p = argv + 2; *p; ++p) { 602 if ((p[0][0] != '1' && p[0][0] != '2') || 603 p[0][1] != '.') 604 break; 605 len = strlen(*p); 606 if (len - 2 != strspn(*p + 2, "0123456789")) 607 break; 608 if ((t = malloc(len + 3)) == NULL) 609 enomem(); 610 t[0] = '-'; 611 t[1] = 'o'; 612 memmove(t + 2, *p, len + 1); 613 *p = t; 614 } 615 argv = p - 1; 616 break; 617 } 618 } 619 } 620 621 void 622 enomem(void) 623 { 624 errx(1, "no memory"); 625 } 626 627 void 628 usage(void) 629 { 630 (void)fprintf(stderr, 631 "usage: %s [-a fileno | -v fileno] [-e string] [-j fileno field]\n" 632 " [-o list] [-t char] [-1 field] [-2 field] file1 file2\n", 633 getprogname()); 634 exit(1); 635 } 636