join.c revision 1.16
1 /* $NetBSD: join.c,v 1.16 1999/08/10 20:09:02 tron 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 1.16 1999/08/10 20:09:02 tron 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 <unistd.h> 62 #include <string.h> 63 #include <unistd.h> 64 65 /* 66 * There's a structure per input file which encapsulates the state of the 67 * file. We repeatedly read lines from each file until we've read in all 68 * the consecutive lines from the file with a common join field. Then we 69 * compare the set of lines with an equivalent set from the other file. 70 */ 71 typedef struct { 72 char *line; /* line */ 73 u_long linealloc; /* line allocated count */ 74 char **fields; /* line field(s) */ 75 u_long fieldcnt; /* line field(s) count */ 76 u_long fieldalloc; /* line field(s) allocated count */ 77 } LINE; 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, 1, 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 __P((LINE *, u_long, LINE *, u_long)); 108 void enomem __P((void)); 109 void fieldarg __P((char *)); 110 void joinlines __P((INPUT *, INPUT *)); 111 int main __P((int, char **)); 112 void obsolete __P((char **)); 113 void outfield __P((LINE *, u_long)); 114 void outoneline __P((INPUT *, LINE *)); 115 void outtwoline __P((INPUT *, LINE *, INPUT *, LINE *)); 116 void slurp __P((INPUT *)); 117 void usage __P((void)); 118 119 int 120 main(argc, argv) 121 int argc; 122 char *argv[]; 123 { 124 INPUT *F1, *F2; 125 int aflag, ch, cval, vflag; 126 char *end; 127 128 F1 = &input1; 129 F2 = &input2; 130 131 aflag = vflag = 0; 132 obsolete(argv); 133 while ((ch = getopt(argc, argv, "\01a:e:j:1:2:o:t:v:")) != -1) { 134 switch (ch) { 135 case '\01': 136 aflag = 1; 137 F1->unpair = F2->unpair = 1; 138 break; 139 case '1': 140 if ((F1->joinf = strtol(optarg, &end, 10)) < 1) { 141 warnx("-1 option field number less than 1"); 142 usage(); 143 } 144 if (*end) { 145 warnx("illegal field number -- %s", optarg); 146 usage(); 147 } 148 --F1->joinf; 149 break; 150 case '2': 151 if ((F2->joinf = strtol(optarg, &end, 10)) < 1) { 152 warnx("-2 option field number less than 1"); 153 usage(); 154 } 155 if (*end) { 156 warnx("illegal field number -- %s", optarg); 157 usage(); 158 } 159 --F2->joinf; 160 break; 161 case 'a': 162 aflag = 1; 163 switch(strtol(optarg, &end, 10)) { 164 case 1: 165 F1->unpair = 1; 166 break; 167 case 2: 168 F2->unpair = 1; 169 break; 170 default: 171 warnx("-a option file number not 1 or 2"); 172 usage(); 173 break; 174 } 175 if (*end) { 176 warnx("illegal file number -- %s", optarg); 177 usage(); 178 } 179 break; 180 case 'e': 181 empty = optarg; 182 break; 183 case 'j': 184 if ((F1->joinf = F2->joinf = 185 strtol(optarg, &end, 10)) < 1) { 186 warnx("-j option field number less than 1"); 187 usage(); 188 } 189 if (*end) { 190 warnx("illegal field number -- %s", optarg); 191 usage(); 192 } 193 --F1->joinf; 194 --F2->joinf; 195 break; 196 case 'o': 197 fieldarg(optarg); 198 break; 199 case 't': 200 spans = 0; 201 if (strlen(tabchar = optarg) != 1) { 202 warnx("illegal tab character specification"); 203 usage(); 204 } 205 break; 206 case 'v': 207 vflag = 1; 208 joinout = 0; 209 switch(strtol(optarg, &end, 10)) { 210 case 1: 211 F1->unpair = 1; 212 break; 213 case 2: 214 F2->unpair = 1; 215 break; 216 default: 217 warnx("-v option file number not 1 or 2"); 218 usage(); 219 break; 220 } 221 if (*end) { 222 warnx("illegal file number -- %s", optarg); 223 usage(); 224 } 225 break; 226 case '?': 227 default: 228 usage(); 229 } 230 } 231 argc -= optind; 232 argv += optind; 233 234 if (aflag && vflag) 235 errx(1, "-a and -v options mutually exclusive"); 236 237 if (argc != 2) 238 usage(); 239 240 /* Open the files; "-" means stdin. */ 241 if (!strcmp(*argv, "-")) 242 F1->fp = stdin; 243 else if ((F1->fp = fopen(*argv, "r")) == NULL) 244 err(1, "%s", *argv); 245 ++argv; 246 if (!strcmp(*argv, "-")) 247 F2->fp = stdin; 248 else if ((F2->fp = fopen(*argv, "r")) == NULL) 249 err(1, "%s", *argv); 250 if (F1->fp == stdin && F2->fp == stdin) 251 errx(1, "only one input file may be stdin"); 252 253 slurp(F1); 254 slurp(F2); 255 while (F1->setcnt && F2->setcnt) { 256 cval = cmp(F1->set, F1->joinf, F2->set, F2->joinf); 257 if (cval == 0) { 258 /* Oh joy, oh rapture, oh beauty divine! */ 259 if (joinout) 260 joinlines(F1, F2); 261 slurp(F1); 262 slurp(F2); 263 } else if (cval < 0) { 264 /* File 1 takes the lead... */ 265 if (F1->unpair) 266 joinlines(F1, NULL); 267 slurp(F1); 268 } else { 269 /* File 2 takes the lead... */ 270 if (F2->unpair) 271 joinlines(F2, NULL); 272 slurp(F2); 273 } 274 } 275 276 /* 277 * Now that one of the files is used up, optionally output any 278 * remaining lines from the other file. 279 */ 280 if (F1->unpair) 281 while (F1->setcnt) { 282 joinlines(F1, NULL); 283 slurp(F1); 284 } 285 if (F2->unpair) 286 while (F2->setcnt) { 287 joinlines(F2, NULL); 288 slurp(F2); 289 } 290 exit(0); 291 } 292 293 void 294 slurp(F) 295 INPUT *F; 296 { 297 LINE *lp, *lastlp; 298 LINE tmp; 299 size_t len; 300 int cnt; 301 char *bp, *fieldp; 302 303 /* 304 * Read all of the lines from an input file that have the same 305 * join field. 306 */ 307 F->setcnt = 0; 308 for (lastlp = NULL;; ++F->setcnt, lastlp = lp) { 309 /* 310 * If we're out of space to hold line structures, allocate 311 * more. Initialize the structure so that we know that this 312 * is new space. 313 */ 314 if (F->setcnt == F->setalloc) { 315 cnt = F->setalloc; 316 F->setalloc += 100; 317 if ((F->set = realloc(F->set, 318 F->setalloc * sizeof(LINE))) == NULL) 319 enomem(); 320 memset(F->set + cnt, 0, 100 * sizeof(LINE)); 321 } 322 323 /* 324 * Get any pushed back line, else get the next line. Allocate 325 * space as necessary. If taking the line from the stack swap 326 * the two structures so that we don't lose the allocated space. 327 * This could be avoided by doing another level of indirection, 328 * but it's probably okay as is. 329 */ 330 lp = &F->set[F->setcnt]; 331 if (F->pushback != -1) { 332 tmp = F->set[F->setcnt]; 333 F->set[F->setcnt] = F->set[F->pushback]; 334 F->set[F->pushback] = tmp; 335 F->pushback = -1; 336 continue; 337 } 338 if ((bp = fgetln(F->fp, &len)) == NULL) 339 return; 340 if (lp->linealloc <= len + 1) { 341 if (lp->linealloc == 0) 342 lp->linealloc = 128; 343 while (lp->linealloc <= len + 1) 344 lp->linealloc *= 2; 345 346 if ((lp->line = realloc(lp->line, 347 lp->linealloc * sizeof(char))) == NULL) 348 enomem(); 349 } 350 memmove(lp->line, bp, len); 351 352 /* Replace trailing newline, if it exists. */ 353 if (bp[len - 1] == '\n') 354 lp->line[len - 1] = '\0'; 355 else 356 lp->line[len] = '\0'; 357 bp = lp->line; 358 359 /* Split the line into fields, allocate space as necessary. */ 360 lp->fieldcnt = 0; 361 while ((fieldp = strsep(&bp, tabchar)) != NULL) { 362 if (spans && *fieldp == '\0') 363 continue; 364 if (lp->fieldcnt == lp->fieldalloc) { 365 lp->fieldalloc += 100; 366 if ((lp->fields = realloc(lp->fields, 367 lp->fieldalloc * sizeof(char *))) == NULL) 368 enomem(); 369 } 370 lp->fields[lp->fieldcnt++] = fieldp; 371 } 372 373 /* See if the join field value has changed. */ 374 if (lastlp != NULL && cmp(lp, F->joinf, lastlp, F->joinf)) { 375 F->pushback = F->setcnt; 376 break; 377 } 378 } 379 } 380 381 int 382 cmp(lp1, fieldno1, lp2, fieldno2) 383 LINE *lp1, *lp2; 384 u_long fieldno1, fieldno2; 385 { 386 387 if (lp1->fieldcnt <= fieldno1) 388 return (lp2->fieldcnt < fieldno2 ? 0 : 1); 389 if (lp2->fieldcnt <= fieldno2) 390 return (-1); 391 return (strcmp(lp1->fields[fieldno1], lp2->fields[fieldno2])); 392 } 393 394 void 395 joinlines(F1, F2) 396 INPUT *F1, *F2; 397 { 398 int cnt1, cnt2; 399 400 /* 401 * Output the results of a join comparison. The output may be from 402 * either file 1 or file 2 (in which case the first argument is the 403 * file from which to output) or from both. 404 */ 405 if (F2 == NULL) { 406 for (cnt1 = 0; cnt1 < F1->setcnt; ++cnt1) 407 outoneline(F1, &F1->set[cnt1]); 408 return; 409 } 410 for (cnt1 = 0; cnt1 < F1->setcnt; ++cnt1) 411 for (cnt2 = 0; cnt2 < F2->setcnt; ++cnt2) 412 outtwoline(F1, &F1->set[cnt1], F2, &F2->set[cnt2]); 413 } 414 415 void 416 outoneline(F, lp) 417 INPUT *F; 418 LINE *lp; 419 { 420 int cnt; 421 422 /* 423 * Output a single line from one of the files, according to the 424 * join rules. This happens when we are writing unmatched single 425 * lines. Output empty fields in the right places. 426 */ 427 if (olist) 428 for (cnt = 0; cnt < olistcnt; ++cnt) { 429 if (olist[cnt].fileno == F->number) 430 outfield(lp, olist[cnt].fieldno); 431 } 432 else 433 for (cnt = 0; cnt < lp->fieldcnt; ++cnt) 434 outfield(lp, cnt); 435 (void)printf("\n"); 436 if (ferror(stdout)) 437 err(1, "stdout"); 438 needsep = 0; 439 } 440 441 void 442 outtwoline(F1, lp1, F2, lp2) 443 INPUT *F1, *F2; 444 LINE *lp1, *lp2; 445 { 446 int cnt; 447 448 /* Output a pair of lines according to the join list (if any). */ 449 if (olist) { 450 for (cnt = 0; cnt < olistcnt; ++cnt) 451 if (olist[cnt].fileno == 1) 452 outfield(lp1, olist[cnt].fieldno); 453 else /* if (olist[cnt].fileno == 2) */ 454 outfield(lp2, olist[cnt].fieldno); 455 } else { 456 /* 457 * Output the join field, then the remaining fields from F1 458 * and F2. 459 */ 460 outfield(lp1, F1->joinf); 461 for (cnt = 0; cnt < lp1->fieldcnt; ++cnt) 462 if (F1->joinf != cnt) 463 outfield(lp1, cnt); 464 for (cnt = 0; cnt < lp2->fieldcnt; ++cnt) 465 if (F2->joinf != cnt) 466 outfield(lp2, cnt); 467 } 468 (void)printf("\n"); 469 if (ferror(stdout)) 470 err(1, "stdout"); 471 needsep = 0; 472 } 473 474 void 475 outfield(lp, fieldno) 476 LINE *lp; 477 u_long fieldno; 478 { 479 if (needsep++) 480 (void)printf("%c", *tabchar); 481 if (!ferror(stdout)) { 482 if (lp->fieldcnt < fieldno) { 483 if (empty != NULL) 484 (void)printf("%s", empty); 485 } else { 486 if (*lp->fields[fieldno] == '\0') 487 return; 488 (void)printf("%s", lp->fields[fieldno]); 489 } 490 } 491 if (ferror(stdout)) 492 err(1, "stdout"); 493 } 494 495 /* 496 * Convert an output list argument "2.1, 1.3, 2.4" into an array of output 497 * fields. 498 */ 499 void 500 fieldarg(option) 501 char *option; 502 { 503 u_long fieldno; 504 char *end, *token; 505 506 while ((token = strsep(&option, " \t")) != NULL) { 507 if (*token == '\0') 508 continue; 509 if ((token[0] != '1' && token[0] != '2') || token[1] != '.') 510 errx(1, "malformed -o option field"); 511 fieldno = strtol(token + 2, &end, 10); 512 if (*end) 513 errx(1, "malformed -o option field"); 514 if (fieldno == 0) 515 errx(1, "field numbers are 1 based"); 516 if (olistcnt == olistalloc) { 517 olistalloc += 50; 518 if ((olist = realloc(olist, 519 olistalloc * sizeof(OLIST))) == NULL) 520 enomem(); 521 } 522 olist[olistcnt].fileno = token[0] - '0'; 523 olist[olistcnt].fieldno = fieldno - 1; 524 ++olistcnt; 525 } 526 } 527 528 void 529 obsolete(argv) 530 char **argv; 531 { 532 int len; 533 char **p, *ap, *t; 534 535 while ((ap = *++argv) != NULL) { 536 /* Return if "--". */ 537 if (ap[0] == '-' && ap[1] == '-') 538 return; 539 switch (ap[1]) { 540 case 'a': 541 /* 542 * The original join allowed "-a", which meant the 543 * same as -a1 plus -a2. POSIX 1003.2, Draft 11.2 544 * only specifies this as "-a 1" and "a -2", so we 545 * have to use another option flag, one that is 546 * unlikely to ever be used or accidentally entered 547 * on the command line. (Well, we could reallocate 548 * the argv array, but that hardly seems worthwhile.) 549 */ 550 if (ap[2] == '\0') 551 ap[1] = '\01'; 552 break; 553 case 'j': 554 /* 555 * The original join allowed "-j[12] arg" and "-j arg". 556 * Convert the former to "-[12] arg". Don't convert 557 * the latter since getopt(3) can handle it. 558 */ 559 switch(ap[2]) { 560 case '1': 561 if (ap[3] != '\0') 562 goto jbad; 563 ap[1] = '1'; 564 ap[2] = '\0'; 565 break; 566 case '2': 567 if (ap[3] != '\0') 568 goto jbad; 569 ap[1] = '2'; 570 ap[2] = '\0'; 571 break; 572 case '\0': 573 break; 574 default: 575 jbad: errx(1, "illegal option -- %s", ap); 576 usage(); 577 } 578 break; 579 case 'o': 580 /* 581 * The original join allowed "-o arg arg". Convert to 582 * "-o arg -o arg". 583 */ 584 if (ap[2] != '\0') 585 break; 586 for (p = argv + 2; *p; ++p) { 587 if ((p[0][0] != '1' && p[0][0] != '2') || 588 p[0][1] != '.') 589 break; 590 len = strlen(*p); 591 if (len - 2 != strspn(*p + 2, "0123456789")) 592 break; 593 if ((t = malloc(len + 3)) == NULL) 594 enomem(); 595 t[0] = '-'; 596 t[1] = 'o'; 597 memmove(t + 2, *p, len + 1); 598 *p = t; 599 } 600 argv = p - 1; 601 break; 602 } 603 } 604 } 605 606 void 607 enomem() 608 { 609 errx(1, "no memory"); 610 } 611 612 void 613 usage() 614 { 615 (void)fprintf(stderr, "%s%s\n", 616 "usage: join [-a fileno | -v fileno ] [-e string] [-1 field] ", 617 "[-2 field]\n [-o list] [-t char] file1 file2"); 618 exit(1); 619 } 620
Indexes created Fri Sep 26 20:09:58 GMT 2025