1 /* $NetBSD: join.c,v 1.25 2004/07/13 11:53:57 wiz 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 #include <sys/cdefs.h> 37 #ifndef lint 38 __COPYRIGHT( 39 "@(#) Copyright (c) 1991 The Regents of the University of California.\n\ 40 All rights reserved.\n"); 41 #endif /* not lint */ 42 43 #ifndef lint 44 #if 0 45 static char sccsid[] = "from: @(#)join.c 5.1 (Berkeley) 11/18/91"; 46 #else 47 __RCSID("$NetBSD: join.c,v 1.25 2004/07/13 11:53:57 wiz Exp $"); 48 #endif 49 #endif /* not lint */ 50 51 #include <sys/types.h> 52 #include <ctype.h> 53 #include <err.h> 54 #include <errno.h> 55 #include <stdio.h> 56 #include <stdlib.h> 57 #include <string.h> 58 #include <unistd.h> 59 60 /* 61 * There's a structure per input file which encapsulates the state of the 62 * file. We repeatedly read lines from each file until we've read in all 63 * the consecutive lines from the file with a common join field. Then we 64 * compare the set of lines with an equivalent set from the other file. 65 */ 66 typedef struct { 67 char *line; /* line */ 68 u_long linealloc; /* line allocated count */ 69 char **fields; /* line field(s) */ 70 u_long fieldcnt; /* line field(s) count */ 71 u_long fieldalloc; /* line field(s) allocated count */ 72 } LINE; 73 74 LINE noline = {"", 0, 0, 0, 0}; /* arg for outfield if no line to output */ 75 76 typedef struct { 77 FILE *fp; /* file descriptor */ 78 u_long joinf; /* join field (-1, -2, -j) */ 79 int unpair; /* output unpairable lines (-a) */ 80 int number; /* 1 for file 1, 2 for file 2 */ 81 82 LINE *set; /* set of lines with same field */ 83 u_long pushback; /* line on the stack */ 84 u_long setcnt; /* set count */ 85 u_long setalloc; /* set allocated count */ 86 } INPUT; 87 INPUT input1 = { NULL, 0, 0, 1, NULL, -1, 0, 0, }, 88 input2 = { NULL, 0, 0, 2, NULL, -1, 0, 0, }; 89 90 typedef struct { 91 u_long fileno; /* file number */ 92 u_long fieldno; /* field number */ 93 } OLIST; 94 OLIST *olist; /* output field list */ 95 u_long olistcnt; /* output field list count */ 96 u_long olistalloc; /* output field allocated count */ 97 98 int joinout = 1; /* show lines with matched join fields (-v) */ 99 int needsep; /* need separator character */ 100 int spans = 1; /* span multiple delimiters (-t) */ 101 char *empty; /* empty field replacement string (-e) */ 102 char *tabchar = " \t"; /* delimiter characters (-t) */ 103 104 int cmp __P((LINE *, u_long, LINE *, u_long)); 105 void enomem __P((void)); 106 void fieldarg __P((char *)); 107 void joinlines __P((INPUT *, INPUT *)); 108 int main __P((int, char **)); 109 void obsolete __P((char **)); 110 void outfield __P((LINE *, u_long)); 111 void outoneline __P((INPUT *, LINE *)); 112 void outtwoline __P((INPUT *, LINE *, INPUT *, LINE *)); 113 void slurp __P((INPUT *)); 114 void usage __P((void)); 115 116 int 117 main(argc, argv) 118 int argc; 119 char *argv[]; 120 { 121 INPUT *F1, *F2; 122 int aflag, ch, cval, vflag; 123 char *end; 124 125 F1 = &input1; 126 F2 = &input2; 127 128 aflag = vflag = 0; 129 obsolete(argv); 130 while ((ch = getopt(argc, argv, "\01a:e:j:1:2:o:t:v:")) != -1) { 131 switch (ch) { 132 case '\01': 133 aflag = 1; 134 F1->unpair = F2->unpair = 1; 135 break; 136 case '1': 137 if ((F1->joinf = strtol(optarg, &end, 10)) < 1) { 138 warnx("-1 option field number less than 1"); 139 usage(); 140 } 141 if (*end) { 142 warnx("illegal field number -- %s", optarg); 143 usage(); 144 } 145 --F1->joinf; 146 break; 147 case '2': 148 if ((F2->joinf = strtol(optarg, &end, 10)) < 1) { 149 warnx("-2 option field number less than 1"); 150 usage(); 151 } 152 if (*end) { 153 warnx("illegal field number -- %s", optarg); 154 usage(); 155 } 156 --F2->joinf; 157 break; 158 case 'a': 159 aflag = 1; 160 switch(strtol(optarg, &end, 10)) { 161 case 1: 162 F1->unpair = 1; 163 break; 164 case 2: 165 F2->unpair = 1; 166 break; 167 default: 168 warnx("-a option file number not 1 or 2"); 169 usage(); 170 break; 171 } 172 if (*end) { 173 warnx("illegal file number -- %s", optarg); 174 usage(); 175 } 176 break; 177 case 'e': 178 empty = optarg; 179 break; 180 case 'j': 181 if ((F1->joinf = F2->joinf = 182 strtol(optarg, &end, 10)) < 1) { 183 warnx("-j option field number less than 1"); 184 usage(); 185 } 186 if (*end) { 187 warnx("illegal field number -- %s", optarg); 188 usage(); 189 } 190 --F1->joinf; 191 --F2->joinf; 192 break; 193 case 'o': 194 fieldarg(optarg); 195 break; 196 case 't': 197 spans = 0; 198 if (strlen(tabchar = optarg) != 1) { 199 warnx("illegal tab character specification"); 200 usage(); 201 } 202 break; 203 case 'v': 204 vflag = 1; 205 joinout = 0; 206 switch(strtol(optarg, &end, 10)) { 207 case 1: 208 F1->unpair = 1; 209 break; 210 case 2: 211 F2->unpair = 1; 212 break; 213 default: 214 warnx("-v option file number not 1 or 2"); 215 usage(); 216 break; 217 } 218 if (*end) { 219 warnx("illegal file number -- %s", optarg); 220 usage(); 221 } 222 break; 223 case '?': 224 default: 225 usage(); 226 } 227 } 228 argc -= optind; 229 argv += optind; 230 231 if (aflag && vflag) 232 errx(1, "-a and -v options mutually exclusive"); 233 234 if (argc != 2) 235 usage(); 236 237 /* Open the files; "-" means stdin. */ 238 if (!strcmp(*argv, "-")) 239 F1->fp = stdin; 240 else if ((F1->fp = fopen(*argv, "r")) == NULL) 241 err(1, "%s", *argv); 242 ++argv; 243 if (!strcmp(*argv, "-")) 244 F2->fp = stdin; 245 else if ((F2->fp = fopen(*argv, "r")) == NULL) 246 err(1, "%s", *argv); 247 if (F1->fp == stdin && F2->fp == stdin) 248 errx(1, "only one input file may be stdin"); 249 250 slurp(F1); 251 slurp(F2); 252 while (F1->setcnt && F2->setcnt) { 253 cval = cmp(F1->set, F1->joinf, F2->set, F2->joinf); 254 if (cval == 0) { 255 /* Oh joy, oh rapture, oh beauty divine! */ 256 if (joinout) 257 joinlines(F1, F2); 258 slurp(F1); 259 slurp(F2); 260 } else if (cval < 0) { 261 /* File 1 takes the lead... */ 262 if (F1->unpair) 263 joinlines(F1, NULL); 264 slurp(F1); 265 } else { 266 /* File 2 takes the lead... */ 267 if (F2->unpair) 268 joinlines(F2, NULL); 269 slurp(F2); 270 } 271 } 272 273 /* 274 * Now that one of the files is used up, optionally output any 275 * remaining lines from the other file. 276 */ 277 if (F1->unpair) 278 while (F1->setcnt) { 279 joinlines(F1, NULL); 280 slurp(F1); 281 } 282 if (F2->unpair) 283 while (F2->setcnt) { 284 joinlines(F2, NULL); 285 slurp(F2); 286 } 287 exit(0); 288 } 289 290 void 291 slurp(F) 292 INPUT *F; 293 { 294 LINE *lp; 295 LINE tmp; 296 LINE *nline; 297 size_t len; 298 int cnt; 299 char *bp, *fieldp; 300 u_long nsize; 301 302 /* 303 * Read all of the lines from an input file that have the same 304 * join field. 305 */ 306 for (F->setcnt = 0;; ++F->setcnt) { 307 /* 308 * If we're out of space to hold line structures, allocate 309 * more. Initialize the structure so that we know that this 310 * is new space. 311 */ 312 if (F->setcnt == F->setalloc) { 313 cnt = F->setalloc; 314 if (F->setalloc == 0) 315 nsize = 64; 316 else 317 nsize = F->setalloc << 1; 318 if ((nline = realloc(F->set, 319 nsize * sizeof(LINE))) == NULL) 320 enomem(); 321 F->set = nline; 322 F->setalloc = nsize; 323 memset(F->set + cnt, 0, 324 (F->setalloc - cnt) * sizeof(LINE)); 325 } 326 327 /* 328 * Get any pushed back line, else get the next line. Allocate 329 * space as necessary. If taking the line from the stack swap 330 * the two structures so that we don't lose the allocated space. 331 * This could be avoided by doing another level of indirection, 332 * but it's probably okay as is. 333 */ 334 lp = &F->set[F->setcnt]; 335 if (F->pushback != -1) { 336 tmp = F->set[F->setcnt]; 337 F->set[F->setcnt] = F->set[F->pushback]; 338 F->set[F->pushback] = tmp; 339 F->pushback = -1; 340 continue; 341 } 342 if ((bp = fgetln(F->fp, &len)) == NULL) 343 return; 344 if (lp->linealloc <= len + 1) { 345 char *n; 346 347 if (lp->linealloc == 0) 348 nsize = 128; 349 else 350 nsize = lp->linealloc; 351 while (nsize <= len + 1) 352 nsize <<= 1; 353 if ((n = realloc(lp->line, 354 nsize * sizeof(char))) == NULL) 355 enomem(); 356 lp->line = n; 357 lp->linealloc = nsize; 358 } 359 memmove(lp->line, bp, len); 360 361 /* Replace trailing newline, if it exists. */ 362 if (bp[len - 1] == '\n') 363 lp->line[len - 1] = '\0'; 364 else 365 lp->line[len] = '\0'; 366 bp = lp->line; 367 368 /* Split the line into fields, allocate space as necessary. */ 369 lp->fieldcnt = 0; 370 while ((fieldp = strsep(&bp, tabchar)) != NULL) { 371 if (spans && *fieldp == '\0') 372 continue; 373 if (lp->fieldcnt == lp->fieldalloc) { 374 char **n; 375 376 if (lp->fieldalloc == 0) 377 nsize = 16; 378 else 379 nsize = lp->fieldalloc << 1; 380 if ((n = realloc(lp->fields, 381 nsize * sizeof(char *))) == NULL) 382 enomem(); 383 lp->fields = n; 384 lp->fieldalloc = nsize; 385 } 386 lp->fields[lp->fieldcnt++] = fieldp; 387 } 388 389 /* See if the join field value has changed. */ 390 if (F->setcnt && cmp(lp, F->joinf, lp - 1, F->joinf)) { 391 F->pushback = F->setcnt; 392 break; 393 } 394 } 395 } 396 397 int 398 cmp(lp1, fieldno1, lp2, fieldno2) 399 LINE *lp1, *lp2; 400 u_long fieldno1, fieldno2; 401 { 402 403 if (lp1->fieldcnt <= fieldno1) 404 return (lp2->fieldcnt <= fieldno2 ? 0 : 1); 405 if (lp2->fieldcnt <= fieldno2) 406 return (-1); 407 return (strcmp(lp1->fields[fieldno1], lp2->fields[fieldno2])); 408 } 409 410 void 411 joinlines(F1, F2) 412 INPUT *F1, *F2; 413 { 414 int cnt1, cnt2; 415 416 /* 417 * Output the results of a join comparison. The output may be from 418 * either file 1 or file 2 (in which case the first argument is the 419 * file from which to output) or from both. 420 */ 421 if (F2 == NULL) { 422 for (cnt1 = 0; cnt1 < F1->setcnt; ++cnt1) 423 outoneline(F1, &F1->set[cnt1]); 424 return; 425 } 426 for (cnt1 = 0; cnt1 < F1->setcnt; ++cnt1) 427 for (cnt2 = 0; cnt2 < F2->setcnt; ++cnt2) 428 outtwoline(F1, &F1->set[cnt1], F2, &F2->set[cnt2]); 429 } 430 431 void 432 outoneline(F, lp) 433 INPUT *F; 434 LINE *lp; 435 { 436 int cnt; 437 438 /* 439 * Output a single line from one of the files, according to the 440 * join rules. This happens when we are writing unmatched single 441 * lines. Output empty fields in the right places. 442 */ 443 if (olist) 444 for (cnt = 0; cnt < olistcnt; ++cnt) { 445 if (olist[cnt].fileno == F->number) 446 outfield(lp, olist[cnt].fieldno); 447 else 448 outfield(&noline, 1); 449 } 450 else 451 for (cnt = 0; cnt < lp->fieldcnt; ++cnt) 452 outfield(lp, cnt); 453 (void)printf("\n"); 454 if (ferror(stdout)) 455 err(1, "stdout"); 456 needsep = 0; 457 } 458 459 void 460 outtwoline(F1, lp1, F2, lp2) 461 INPUT *F1, *F2; 462 LINE *lp1, *lp2; 463 { 464 int cnt; 465 466 /* Output a pair of lines according to the join list (if any). */ 467 if (olist) { 468 for (cnt = 0; cnt < olistcnt; ++cnt) 469 if (olist[cnt].fileno == 1) 470 outfield(lp1, olist[cnt].fieldno); 471 else /* if (olist[cnt].fileno == 2) */ 472 outfield(lp2, olist[cnt].fieldno); 473 } else { 474 /* 475 * Output the join field, then the remaining fields from F1 476 * and F2. 477 */ 478 outfield(lp1, F1->joinf); 479 for (cnt = 0; cnt < lp1->fieldcnt; ++cnt) 480 if (F1->joinf != cnt) 481 outfield(lp1, cnt); 482 for (cnt = 0; cnt < lp2->fieldcnt; ++cnt) 483 if (F2->joinf != cnt) 484 outfield(lp2, cnt); 485 } 486 (void)printf("\n"); 487 if (ferror(stdout)) 488 err(1, "stdout"); 489 needsep = 0; 490 } 491 492 void 493 outfield(lp, fieldno) 494 LINE *lp; 495 u_long fieldno; 496 { 497 if (needsep++) 498 (void)printf("%c", *tabchar); 499 if (!ferror(stdout)) { 500 if (lp->fieldcnt <= fieldno) { 501 if (empty != NULL) 502 (void)printf("%s", empty); 503 } else { 504 if (*lp->fields[fieldno] == '\0') 505 return; 506 (void)printf("%s", lp->fields[fieldno]); 507 } 508 } 509 if (ferror(stdout)) 510 err(1, "stdout"); 511 } 512 513 /* 514 * Convert an output list argument "2.1, 1.3, 2.4" into an array of output 515 * fields. 516 */ 517 void 518 fieldarg(option) 519 char *option; 520 { 521 u_long fieldno; 522 char *end, *token; 523 OLIST *n; 524 525 while ((token = strsep(&option, ", \t")) != NULL) { 526 if (*token == '\0') 527 continue; 528 if ((token[0] != '1' && token[0] != '2') || token[1] != '.') 529 errx(1, "malformed -o option field"); 530 fieldno = strtol(token + 2, &end, 10); 531 if (*end) 532 errx(1, "malformed -o option field"); 533 if (fieldno == 0) 534 errx(1, "field numbers are 1 based"); 535 if (olistcnt == olistalloc) { 536 if ((n = realloc(olist, 537 (olistalloc + 50) * sizeof(OLIST))) == NULL) 538 enomem(); 539 olist = n; 540 olistalloc += 50; 541 } 542 olist[olistcnt].fileno = token[0] - '0'; 543 olist[olistcnt].fieldno = fieldno - 1; 544 ++olistcnt; 545 } 546 } 547 548 void 549 obsolete(argv) 550 char **argv; 551 { 552 int len; 553 char **p, *ap, *t; 554 555 while ((ap = *++argv) != NULL) { 556 /* Return if "--". */ 557 if (ap[0] == '-' && ap[1] == '-') 558 return; 559 switch (ap[1]) { 560 case 'a': 561 /* 562 * The original join allowed "-a", which meant the 563 * same as -a1 plus -a2. POSIX 1003.2, Draft 11.2 564 * only specifies this as "-a 1" and "a -2", so we 565 * have to use another option flag, one that is 566 * unlikely to ever be used or accidentally entered 567 * on the command line. (Well, we could reallocate 568 * the argv array, but that hardly seems worthwhile.) 569 */ 570 if (ap[2] == '\0') 571 ap[1] = '\01'; 572 break; 573 case 'j': 574 /* 575 * The original join allowed "-j[12] arg" and "-j arg". 576 * Convert the former to "-[12] arg". Don't convert 577 * the latter since getopt(3) can handle it. 578 */ 579 switch(ap[2]) { 580 case '1': 581 if (ap[3] != '\0') 582 goto jbad; 583 ap[1] = '1'; 584 ap[2] = '\0'; 585 break; 586 case '2': 587 if (ap[3] != '\0') 588 goto jbad; 589 ap[1] = '2'; 590 ap[2] = '\0'; 591 break; 592 case '\0': 593 break; 594 default: 595 jbad: errx(1, "illegal option -- %s", ap); 596 usage(); 597 } 598 break; 599 case 'o': 600 /* 601 * The original join allowed "-o arg arg". Convert to 602 * "-o arg -o arg". 603 */ 604 if (ap[2] != '\0') 605 break; 606 for (p = argv + 2; *p; ++p) { 607 if ((p[0][0] != '1' && p[0][0] != '2') || 608 p[0][1] != '.') 609 break; 610 len = strlen(*p); 611 if (len - 2 != strspn(*p + 2, "0123456789")) 612 break; 613 if ((t = malloc(len + 3)) == NULL) 614 enomem(); 615 t[0] = '-'; 616 t[1] = 'o'; 617 memmove(t + 2, *p, len + 1); 618 *p = t; 619 } 620 argv = p - 1; 621 break; 622 } 623 } 624 } 625 626 void 627 enomem() 628 { 629 errx(1, "no memory"); 630 } 631 632 void 633 usage() 634 { 635 (void)fprintf(stderr, 636 "usage: %s [-a fileno | -v fileno] [-e string] [-j fileno field]\n" 637 " [-o list] [-t char] [-1 field] [-2 field] file1 file2\n", 638 getprogname()); 639 exit(1); 640 } 641