join.c revision 1.13
1 /* $NetBSD: join.c,v 1.13 1998/08/25 20:59:38 ross 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. All advertising materials mentioning features or use of this software 20 * must display the following acknowledgement: 21 * This product includes software developed by the University of 22 * California, Berkeley and its contributors. 23 * 4. Neither the name of the University nor the names of its contributors 24 * may be used to endorse or promote products derived from this software 25 * without specific prior written permission. 26 * 27 * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND 28 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE 29 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE 30 * ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE 31 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL 32 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS 33 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) 34 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT 35 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY 36 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF 37 * SUCH DAMAGE. 38 */ 39 40 #include <sys/cdefs.h> 41 #ifndef lint 42 __COPYRIGHT( 43 "@(#) Copyright (c) 1991 The Regents of the University of California.\n\ 44 All rights reserved.\n"); 45 #endif /* not lint */ 46 47 #ifndef lint 48 #if 0 49 static char sccsid[] = "from: @(#)join.c 5.1 (Berkeley) 11/18/91"; 50 #else 51 __RCSID("$NetBSD: join.c,v 1.13 1998/08/25 20:59:38 ross Exp $"); 52 #endif 53 #endif /* not lint */ 54 55 #include <sys/types.h> 56 #include <ctype.h> 57 #include <err.h> 58 #include <errno.h> 59 #include <stdio.h> 60 #include <stdlib.h> 61 #include <string.h> 62 #include <unistd.h> 63 64 /* 65 * There's a structure per input file which encapsulates the state of the 66 * file. We repeatedly read lines from each file until we've read in all 67 * the consecutive lines from the file with a common join field. Then we 68 * compare the set of lines with an equivalent set from the other file. 69 */ 70 typedef struct { 71 char *line; /* line */ 72 u_long linealloc; /* line allocated count */ 73 char **fields; /* line field(s) */ 74 u_long fieldcnt; /* line field(s) count */ 75 u_long fieldalloc; /* line field(s) allocated count */ 76 } LINE; 77 78 typedef struct { 79 FILE *fp; /* file descriptor */ 80 u_long joinf; /* join field (-1, -2, -j) */ 81 int unpair; /* output unpairable lines (-a) */ 82 int number; /* 1 for file 1, 2 for file 2 */ 83 84 LINE *set; /* set of lines with same field */ 85 u_long pushback; /* line on the stack */ 86 u_long setcnt; /* set count */ 87 u_long setalloc; /* set allocated count */ 88 } INPUT; 89 INPUT input1 = { NULL, 0, 0, 1, NULL, -1, 0, 0, }, 90 input2 = { NULL, 0, 0, 1, NULL, -1, 0, 0, }; 91 92 typedef struct { 93 u_long fileno; /* file number */ 94 u_long fieldno; /* field number */ 95 } OLIST; 96 OLIST *olist; /* output field list */ 97 u_long olistcnt; /* output field list count */ 98 u_long olistalloc; /* output field allocated count */ 99 100 int joinout = 1; /* show lines with matched join fields (-v) */ 101 int needsep; /* need separator character */ 102 int spans = 1; /* span multiple delimiters (-t) */ 103 char *empty; /* empty field replacement string (-e) */ 104 char *tabchar = " \t"; /* delimiter characters (-t) */ 105 106 int cmp __P((LINE *, u_long, LINE *, u_long)); 107 void enomem __P((void)); 108 void fieldarg __P((char *)); 109 void joinlines __P((INPUT *, INPUT *)); 110 int main __P((int, char **)); 111 void obsolete __P((char **)); 112 void outfield __P((LINE *, u_long)); 113 void outoneline __P((INPUT *, LINE *)); 114 void outtwoline __P((INPUT *, LINE *, INPUT *, LINE *)); 115 void slurp __P((INPUT *)); 116 void usage __P((void)); 117 118 int 119 main(argc, argv) 120 int argc; 121 char *argv[]; 122 { 123 INPUT *F1, *F2; 124 int aflag, ch, cval, vflag; 125 char *end; 126 127 F1 = &input1; 128 F2 = &input2; 129 130 aflag = vflag = 0; 131 obsolete(argv); 132 while ((ch = getopt(argc, argv, "\01a:e:j:1:2:o:t:v:")) != -1) { 133 switch (ch) { 134 case '\01': 135 aflag = 1; 136 F1->unpair = F2->unpair = 1; 137 break; 138 case '1': 139 if ((F1->joinf = strtol(optarg, &end, 10)) < 1) { 140 warnx("-1 option field number less than 1"); 141 usage(); 142 } 143 if (*end) { 144 warnx("illegal field number -- %s", optarg); 145 usage(); 146 } 147 --F1->joinf; 148 break; 149 case '2': 150 if ((F2->joinf = strtol(optarg, &end, 10)) < 1) { 151 warnx("-2 option field number less than 1"); 152 usage(); 153 } 154 if (*end) { 155 warnx("illegal field number -- %s", optarg); 156 usage(); 157 } 158 --F2->joinf; 159 break; 160 case 'a': 161 aflag = 1; 162 switch(strtol(optarg, &end, 10)) { 163 case 1: 164 F1->unpair = 1; 165 break; 166 case 2: 167 F2->unpair = 1; 168 break; 169 default: 170 warnx("-a option file number not 1 or 2"); 171 usage(); 172 break; 173 } 174 if (*end) { 175 warnx("illegal file number -- %s", optarg); 176 usage(); 177 } 178 break; 179 case 'e': 180 empty = optarg; 181 break; 182 case 'j': 183 if ((F1->joinf = F2->joinf = 184 strtol(optarg, &end, 10)) < 1) { 185 warnx("-j option field number less than 1"); 186 usage(); 187 } 188 if (*end) { 189 warnx("illegal field number -- %s", optarg); 190 usage(); 191 } 192 --F1->joinf; 193 --F2->joinf; 194 break; 195 case 'o': 196 fieldarg(optarg); 197 break; 198 case 't': 199 spans = 0; 200 if (strlen(tabchar = optarg) != 1) { 201 warnx("illegal tab character specification"); 202 usage(); 203 } 204 break; 205 case 'v': 206 vflag = 1; 207 joinout = 0; 208 switch(strtol(optarg, &end, 10)) { 209 case 1: 210 F1->unpair = 1; 211 break; 212 case 2: 213 F2->unpair = 1; 214 break; 215 default: 216 warnx("-v option file number not 1 or 2"); 217 usage(); 218 break; 219 } 220 if (*end) { 221 warnx("illegal file number -- %s", optarg); 222 usage(); 223 } 224 break; 225 case '?': 226 default: 227 usage(); 228 } 229 } 230 argc -= optind; 231 argv += optind; 232 233 if (aflag && vflag) 234 errx(1, "-a and -v options mutually exclusive"); 235 236 if (argc != 2) 237 usage(); 238 239 /* Open the files; "-" means stdin. */ 240 if (!strcmp(*argv, "-")) 241 F1->fp = stdin; 242 else if ((F1->fp = fopen(*argv, "r")) == NULL) 243 err(1, "%s", *argv); 244 ++argv; 245 if (!strcmp(*argv, "-")) 246 F2->fp = stdin; 247 else if ((F2->fp = fopen(*argv, "r")) == NULL) 248 err(1, "%s", *argv); 249 if (F1->fp == stdin && F2->fp == stdin) 250 errx(1, "only one input file may be stdin"); 251 252 slurp(F1); 253 slurp(F2); 254 while (F1->setcnt && F2->setcnt) { 255 cval = cmp(F1->set, F1->joinf, F2->set, F2->joinf); 256 if (cval == 0) { 257 /* Oh joy, oh rapture, oh beauty divine! */ 258 if (joinout) 259 joinlines(F1, F2); 260 slurp(F1); 261 slurp(F2); 262 } else if (cval < 0) { 263 /* File 1 takes the lead... */ 264 if (F1->unpair) 265 joinlines(F1, NULL); 266 slurp(F1); 267 } else { 268 /* File 2 takes the lead... */ 269 if (F2->unpair) 270 joinlines(F2, NULL); 271 slurp(F2); 272 } 273 } 274 275 /* 276 * Now that one of the files is used up, optionally output any 277 * remaining lines from the other file. 278 */ 279 if (F1->unpair) 280 while (F1->setcnt) { 281 joinlines(F1, NULL); 282 slurp(F1); 283 } 284 if (F2->unpair) 285 while (F2->setcnt) { 286 joinlines(F2, NULL); 287 slurp(F2); 288 } 289 exit(0); 290 } 291 292 void 293 slurp(F) 294 INPUT *F; 295 { 296 LINE *lp, *lastlp; 297 LINE tmp; 298 size_t len; 299 int cnt; 300 char *bp, *fieldp; 301 302 /* 303 * Read all of the lines from an input file that have the same 304 * join field. 305 */ 306 F->setcnt = 0; 307 for (lastlp = NULL;; ++F->setcnt, lastlp = lp) { 308 /* 309 * If we're out of space to hold line structures, allocate 310 * more. Initialize the structure so that we know that this 311 * is new space. 312 */ 313 if (F->setcnt == F->setalloc) { 314 cnt = F->setalloc; 315 F->setalloc += 100; 316 if ((F->set = realloc(F->set, 317 F->setalloc * sizeof(LINE))) == NULL) 318 enomem(); 319 memset(F->set + cnt, 0, 100 * sizeof(LINE *)); 320 } 321 322 /* 323 * Get any pushed back line, else get the next line. Allocate 324 * space as necessary. If taking the line from the stack swap 325 * the two structures so that we don't lose the allocated space. 326 * This could be avoided by doing another level of indirection, 327 * but it's probably okay as is. 328 */ 329 lp = &F->set[F->setcnt]; 330 if (F->pushback != -1) { 331 tmp = F->set[F->setcnt]; 332 F->set[F->setcnt] = F->set[F->pushback]; 333 F->set[F->pushback] = tmp; 334 F->pushback = -1; 335 continue; 336 } 337 if ((bp = fgetln(F->fp, &len)) == NULL) 338 return; 339 if (lp->linealloc <= len + 1) { 340 if (lp->linealloc == 0) 341 lp->linealloc = 128; 342 while (lp->linealloc <= len + 1) 343 lp->linealloc *= 2; 344 345 if ((lp->line = realloc(lp->line, 346 lp->linealloc * sizeof(char))) == NULL) 347 enomem(); 348 } 349 memmove(lp->line, bp, len+1); 350 351 /* Replace trailing newline, if it exists. */ 352 if (bp[len - 1] == '\n') 353 lp->line[len - 1] = '\0'; 354 else 355 lp->line[len] = '\0'; 356 bp = lp->line; 357 358 /* Split the line into fields, allocate space as necessary. */ 359 lp->fieldcnt = 0; 360 while ((fieldp = strsep(&bp, tabchar)) != NULL) { 361 if (spans && *fieldp == '\0') 362 continue; 363 if (lp->fieldcnt == lp->fieldalloc) { 364 lp->fieldalloc += 100; 365 if ((lp->fields = realloc(lp->fields, 366 lp->fieldalloc * sizeof(char *))) == NULL) 367 enomem(); 368 } 369 lp->fields[lp->fieldcnt++] = fieldp; 370 } 371 372 /* See if the join field value has changed. */ 373 if (lastlp != NULL && cmp(lp, F->joinf, lastlp, F->joinf)) { 374 F->pushback = F->setcnt; 375 break; 376 } 377 } 378 } 379 380 int 381 cmp(lp1, fieldno1, lp2, fieldno2) 382 LINE *lp1, *lp2; 383 u_long fieldno1, fieldno2; 384 { 385 386 if (lp1->fieldcnt <= fieldno1) 387 return (lp2->fieldcnt < fieldno2 ? 0 : 1); 388 if (lp2->fieldcnt <= fieldno2) 389 return (-1); 390 return (strcmp(lp1->fields[fieldno1], lp2->fields[fieldno2])); 391 } 392 393 void 394 joinlines(F1, F2) 395 INPUT *F1, *F2; 396 { 397 int cnt1, cnt2; 398 399 /* 400 * Output the results of a join comparison. The output may be from 401 * either file 1 or file 2 (in which case the first argument is the 402 * file from which to output) or from both. 403 */ 404 if (F2 == NULL) { 405 for (cnt1 = 0; cnt1 < F1->setcnt; ++cnt1) 406 outoneline(F1, &F1->set[cnt1]); 407 return; 408 } 409 for (cnt1 = 0; cnt1 < F1->setcnt; ++cnt1) 410 for (cnt2 = 0; cnt2 < F2->setcnt; ++cnt2) 411 outtwoline(F1, &F1->set[cnt1], F2, &F2->set[cnt2]); 412 } 413 414 void 415 outoneline(F, lp) 416 INPUT *F; 417 LINE *lp; 418 { 419 int cnt; 420 421 /* 422 * Output a single line from one of the files, according to the 423 * join rules. This happens when we are writing unmatched single 424 * lines. Output empty fields in the right places. 425 */ 426 if (olist) 427 for (cnt = 0; cnt < olistcnt; ++cnt) { 428 if (olist[cnt].fileno == F->number) 429 outfield(lp, olist[cnt].fieldno); 430 } 431 else 432 for (cnt = 0; cnt < lp->fieldcnt; ++cnt) 433 outfield(lp, cnt); 434 (void)printf("\n"); 435 if (ferror(stdout)) 436 err(1, "stdout"); 437 needsep = 0; 438 } 439 440 void 441 outtwoline(F1, lp1, F2, lp2) 442 INPUT *F1, *F2; 443 LINE *lp1, *lp2; 444 { 445 int cnt; 446 447 /* Output a pair of lines according to the join list (if any). */ 448 if (olist) 449 for (cnt = 0; cnt < olistcnt; ++cnt) 450 if (olist[cnt].fileno == 1) 451 outfield(lp1, olist[cnt].fieldno); 452 else /* if (olist[cnt].fileno == 2) */ 453 outfield(lp2, olist[cnt].fieldno); 454 else { 455 /* 456 * Output the join field, then the remaining fields from F1 457 * and F2. 458 */ 459 outfield(lp1, F1->joinf); 460 for (cnt = 0; cnt < lp1->fieldcnt; ++cnt) 461 if (F1->joinf != cnt) 462 outfield(lp1, cnt); 463 for (cnt = 0; cnt < lp2->fieldcnt; ++cnt) 464 if (F2->joinf != cnt) 465 outfield(lp2, cnt); 466 } 467 (void)printf("\n"); 468 if (ferror(stdout)) 469 err(1, "stdout"); 470 needsep = 0; 471 } 472 473 void 474 outfield(lp, fieldno) 475 LINE *lp; 476 u_long fieldno; 477 { 478 if (needsep++) 479 (void)printf("%c", *tabchar); 480 if (!ferror(stdout)) { 481 if (lp->fieldcnt < fieldno) { 482 if (empty != NULL) 483 (void)printf("%s", empty); 484 } else { 485 if (*lp->fields[fieldno] == '\0') 486 return; 487 (void)printf("%s", lp->fields[fieldno]); 488 } 489 } 490 if (ferror(stdout)) 491 err(1, "stdout"); 492 } 493 494 /* 495 * Convert an output list argument "2.1, 1.3, 2.4" into an array of output 496 * fields. 497 */ 498 void 499 fieldarg(option) 500 char *option; 501 { 502 u_long fieldno; 503 char *end, *token; 504 505 while ((token = strsep(&option, " \t")) != NULL) { 506 if (*token == '\0') 507 continue; 508 if ((token[0] != '1' && token[0] != '2') || token[1] != '.') 509 errx(1, "malformed -o option field"); 510 fieldno = strtol(token + 2, &end, 10); 511 if (*end) 512 errx(1, "malformed -o option field"); 513 if (fieldno == 0) 514 errx(1, "field numbers are 1 based"); 515 if (olistcnt == olistalloc) { 516 olistalloc += 50; 517 if ((olist = realloc(olist, 518 olistalloc * sizeof(OLIST))) == NULL) 519 enomem(); 520 } 521 olist[olistcnt].fileno = token[0] - '0'; 522 olist[olistcnt].fieldno = fieldno - 1; 523 ++olistcnt; 524 } 525 } 526 527 void 528 obsolete(argv) 529 char **argv; 530 { 531 int len; 532 char **p, *ap, *t; 533 534 while ((ap = *++argv) != NULL) { 535 /* Return if "--". */ 536 if (ap[0] == '-' && ap[1] == '-') 537 return; 538 switch (ap[1]) { 539 case 'a': 540 /* 541 * The original join allowed "-a", which meant the 542 * same as -a1 plus -a2. POSIX 1003.2, Draft 11.2 543 * only specifies this as "-a 1" and "a -2", so we 544 * have to use another option flag, one that is 545 * unlikely to ever be used or accidentally entered 546 * on the command line. (Well, we could reallocate 547 * the argv array, but that hardly seems worthwhile.) 548 */ 549 if (ap[2] == '\0') 550 ap[1] = '\01'; 551 break; 552 case 'j': 553 /* 554 * The original join allowed "-j[12] arg" and "-j arg". 555 * Convert the former to "-[12] arg". Don't convert 556 * the latter since getopt(3) can handle it. 557 */ 558 switch(ap[2]) { 559 case '1': 560 if (ap[3] != '\0') 561 goto jbad; 562 ap[1] = '1'; 563 ap[2] = '\0'; 564 break; 565 case '2': 566 if (ap[3] != '\0') 567 goto jbad; 568 ap[1] = '2'; 569 ap[2] = '\0'; 570 break; 571 case '\0': 572 break; 573 default: 574 jbad: errx(1, "illegal option -- %s", ap); 575 usage(); 576 } 577 break; 578 case 'o': 579 /* 580 * The original join allowed "-o arg arg". Convert to 581 * "-o arg -o arg". 582 */ 583 if (ap[2] != '\0') 584 break; 585 for (p = argv + 2; *p; ++p) { 586 if ((p[0][0] != '1' && p[0][0] != '2') || 587 p[0][1] != '.') 588 break; 589 len = strlen(*p); 590 if (len - 2 != strspn(*p + 2, "0123456789")) 591 break; 592 if ((t = malloc(len + 3)) == NULL) 593 enomem(); 594 t[0] = '-'; 595 t[1] = 'o'; 596 memmove(t + 2, *p, len + 1); 597 *p = t; 598 } 599 argv = p - 1; 600 break; 601 } 602 } 603 } 604 605 void 606 enomem() 607 { 608 errx(1, "no memory"); 609 } 610 611 void 612 usage() 613 { 614 (void)fprintf(stderr, "%s%s\n", 615 "usage: join [-a fileno | -v fileno ] [-e string] [-1 field] ", 616 "[-2 field]\n [-o list] [-t char] file1 file2"); 617 exit(1); 618 } 619
Indexes created Thu Sep 25 08:09:54 GMT 2025