join.c revision 1.11
1 /* $NetBSD: join.c,v 1.11 1997/10/19 03:32:13 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. 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 1.11 1997/10/19 03:32:13 lukem 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 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 typedef struct { 78 FILE *fp; /* file descriptor */ 79 u_long joinf; /* join field (-1, -2, -j) */ 80 int unpair; /* output unpairable lines (-a) */ 81 int number; /* 1 for file 1, 2 for file 2 */ 82 83 LINE *set; /* set of lines with same field */ 84 u_long pushback; /* line on the stack */ 85 u_long setcnt; /* set count */ 86 u_long setalloc; /* set allocated count */ 87 } INPUT; 88 INPUT input1 = { NULL, 0, 0, 1, NULL, -1, 0, 0, }, 89 input2 = { NULL, 0, 0, 1, NULL, -1, 0, 0, }; 90 91 typedef struct { 92 u_long fileno; /* file number */ 93 u_long fieldno; /* field number */ 94 } OLIST; 95 OLIST *olist; /* output field list */ 96 u_long olistcnt; /* output field list count */ 97 u_long olistalloc; /* output field allocated count */ 98 99 int joinout = 1; /* show lines with matched join fields (-v) */ 100 int needsep; /* need separator character */ 101 int spans = 1; /* span multiple delimiters (-t) */ 102 char *empty; /* empty field replacement string (-e) */ 103 char *tabchar = " \t"; /* delimiter characters (-t) */ 104 105 int cmp __P((LINE *, u_long, LINE *, u_long)); 106 void enomem __P((void)); 107 void fieldarg __P((char *)); 108 void joinlines __P((INPUT *, INPUT *)); 109 int main __P((int, char **)); 110 void obsolete __P((char **)); 111 void outfield __P((LINE *, u_long)); 112 void outoneline __P((INPUT *, LINE *)); 113 void outtwoline __P((INPUT *, LINE *, INPUT *, LINE *)); 114 void slurp __P((INPUT *)); 115 void usage __P((void)); 116 117 int 118 main(argc, argv) 119 int argc; 120 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 (F2->unpair) 284 while (F2->setcnt) { 285 joinlines(F2, NULL); 286 slurp(F2); 287 } 288 exit(0); 289 } 290 291 void 292 slurp(F) 293 INPUT *F; 294 { 295 LINE *lp, *lastlp; 296 LINE tmp; 297 size_t len; 298 int cnt; 299 char *bp, *fieldp; 300 301 /* 302 * Read all of the lines from an input file that have the same 303 * join field. 304 */ 305 F->setcnt = 0; 306 for (lastlp = NULL;; ++F->setcnt, lastlp = lp) { 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 F->setalloc += 100; 315 if ((F->set = realloc(F->set, 316 F->setalloc * sizeof(LINE))) == NULL) 317 enomem(); 318 memset(F->set + cnt, 0, 100 * sizeof(LINE *)); 319 } 320 321 /* 322 * Get any pushed back line, else get the next line. Allocate 323 * space as necessary. If taking the line from the stack swap 324 * the two structures so that we don't lose the allocated space. 325 * This could be avoided by doing another level of indirection, 326 * but it's probably okay as is. 327 */ 328 lp = &F->set[F->setcnt]; 329 if (F->pushback != -1) { 330 tmp = F->set[F->setcnt]; 331 F->set[F->setcnt] = F->set[F->pushback]; 332 F->set[F->pushback] = tmp; 333 F->pushback = -1; 334 continue; 335 } 336 if ((bp = fgetln(F->fp, &len)) == NULL) 337 return; 338 if (lp->linealloc <= len + 1) { 339 if (lp->linealloc == 0) 340 lp->linealloc = 128; 341 while (lp->linealloc <= len + 1) 342 lp->linealloc *= 2; 343 344 if ((lp->line = realloc(lp->line, 345 lp->linealloc * sizeof(char))) == NULL) 346 enomem(); 347 } 348 memmove(lp->line, bp, len+1); 349 350 /* Replace trailing newline, if it exists. */ 351 if (bp[len - 1] == '\n') 352 lp->line[len - 1] = '\0'; 353 else 354 lp->line[len] = '\0'; 355 bp = lp->line; 356 357 /* Split the line into fields, allocate space as necessary. */ 358 lp->fieldcnt = 0; 359 while ((fieldp = strsep(&bp, tabchar)) != NULL) { 360 if (spans && *fieldp == '\0') 361 continue; 362 if (lp->fieldcnt == lp->fieldalloc) { 363 lp->fieldalloc += 100; 364 if ((lp->fields = realloc(lp->fields, 365 lp->fieldalloc * sizeof(char *))) == NULL) 366 enomem(); 367 } 368 lp->fields[lp->fieldcnt++] = fieldp; 369 } 370 371 /* See if the join field value has changed. */ 372 if (lastlp != NULL && cmp(lp, F->joinf, lastlp, F->joinf)) { 373 F->pushback = F->setcnt; 374 break; 375 } 376 } 377 } 378 379 int 380 cmp(lp1, fieldno1, lp2, fieldno2) 381 LINE *lp1, *lp2; 382 u_long fieldno1, fieldno2; 383 { 384 385 if (lp1->fieldcnt <= fieldno1) 386 return (lp2->fieldcnt < fieldno2 ? 0 : 1); 387 if (lp2->fieldcnt <= fieldno2) 388 return (-1); 389 return (strcmp(lp1->fields[fieldno1], lp2->fields[fieldno2])); 390 } 391 392 void 393 joinlines(F1, F2) 394 INPUT *F1, *F2; 395 { 396 int cnt1, cnt2; 397 398 /* 399 * Output the results of a join comparison. The output may be from 400 * either file 1 or file 2 (in which case the first argument is the 401 * file from which to output) or from both. 402 */ 403 if (F2 == NULL) { 404 for (cnt1 = 0; cnt1 < F1->setcnt; ++cnt1) 405 outoneline(F1, &F1->set[cnt1]); 406 return; 407 } 408 for (cnt1 = 0; cnt1 < F1->setcnt; ++cnt1) 409 for (cnt2 = 0; cnt2 < F2->setcnt; ++cnt2) 410 outtwoline(F1, &F1->set[cnt1], F2, &F2->set[cnt2]); 411 } 412 413 void 414 outoneline(F, lp) 415 INPUT *F; 416 LINE *lp; 417 { 418 int cnt; 419 420 /* 421 * Output a single line from one of the files, according to the 422 * join rules. This happens when we are writing unmatched single 423 * lines. Output empty fields in the right places. 424 */ 425 if (olist) 426 for (cnt = 0; cnt < olistcnt; ++cnt) { 427 if (olist[cnt].fileno == F->number) 428 outfield(lp, olist[cnt].fieldno); 429 } 430 else 431 for (cnt = 0; cnt < lp->fieldcnt; ++cnt) 432 outfield(lp, cnt); 433 (void)printf("\n"); 434 if (ferror(stdout)) 435 err(1, "stdout"); 436 needsep = 0; 437 } 438 439 void 440 outtwoline(F1, lp1, F2, lp2) 441 INPUT *F1, *F2; 442 LINE *lp1, *lp2; 443 { 444 int cnt; 445 446 /* Output a pair of lines according to the join list (if any). */ 447 if (olist) 448 for (cnt = 0; cnt < olistcnt; ++cnt) 449 if (olist[cnt].fileno == 1) 450 outfield(lp1, olist[cnt].fieldno); 451 else /* if (olist[cnt].fileno == 2) */ 452 outfield(lp2, olist[cnt].fieldno); 453 else { 454 /* 455 * Output the join field, then the remaining fields from F1 456 * and F2. 457 */ 458 outfield(lp1, F1->joinf); 459 for (cnt = 0; cnt < lp1->fieldcnt; ++cnt) 460 if (F1->joinf != cnt) 461 outfield(lp1, cnt); 462 for (cnt = 0; cnt < lp2->fieldcnt; ++cnt) 463 if (F2->joinf != cnt) 464 outfield(lp2, cnt); 465 } 466 (void)printf("\n"); 467 if (ferror(stdout)) 468 err(1, "stdout"); 469 needsep = 0; 470 } 471 472 void 473 outfield(lp, fieldno) 474 LINE *lp; 475 u_long fieldno; 476 { 477 if (needsep++) 478 (void)printf("%c", *tabchar); 479 if (!ferror(stdout)) 480 if (lp->fieldcnt < fieldno) { 481 if (empty != NULL) 482 (void)printf("%s", empty); 483 } else { 484 if (*lp->fields[fieldno] == '\0') 485 return; 486 (void)printf("%s", lp->fields[fieldno]); 487 } 488 if (ferror(stdout)) 489 err(1, "stdout"); 490 } 491 492 /* 493 * Convert an output list argument "2.1, 1.3, 2.4" into an array of output 494 * fields. 495 */ 496 void 497 fieldarg(option) 498 char *option; 499 { 500 u_long fieldno; 501 char *end, *token; 502 503 while ((token = strsep(&option, " \t")) != NULL) { 504 if (*token == '\0') 505 continue; 506 if ((token[0] != '1' && token[0] != '2') || token[1] != '.') 507 errx(1, "malformed -o option field"); 508 fieldno = strtol(token + 2, &end, 10); 509 if (*end) 510 errx(1, "malformed -o option field"); 511 if (fieldno == 0) 512 errx(1, "field numbers are 1 based"); 513 if (olistcnt == olistalloc) { 514 olistalloc += 50; 515 if ((olist = realloc(olist, 516 olistalloc * sizeof(OLIST))) == NULL) 517 enomem(); 518 } 519 olist[olistcnt].fileno = token[0] - '0'; 520 olist[olistcnt].fieldno = fieldno - 1; 521 ++olistcnt; 522 } 523 } 524 525 void 526 obsolete(argv) 527 char **argv; 528 { 529 int len; 530 char **p, *ap, *t; 531 532 while ((ap = *++argv) != NULL) { 533 /* Return if "--". */ 534 if (ap[0] == '-' && ap[1] == '-') 535 return; 536 switch (ap[1]) { 537 case 'a': 538 /* 539 * The original join allowed "-a", which meant the 540 * same as -a1 plus -a2. POSIX 1003.2, Draft 11.2 541 * only specifies this as "-a 1" and "a -2", so we 542 * have to use another option flag, one that is 543 * unlikely to ever be used or accidentally entered 544 * on the command line. (Well, we could reallocate 545 * the argv array, but that hardly seems worthwhile.) 546 */ 547 if (ap[2] == '\0') 548 ap[1] = '\01'; 549 break; 550 case 'j': 551 /* 552 * The original join allowed "-j[12] arg" and "-j arg". 553 * Convert the former to "-[12] arg". Don't convert 554 * the latter since getopt(3) can handle it. 555 */ 556 switch(ap[2]) { 557 case '1': 558 if (ap[3] != '\0') 559 goto jbad; 560 ap[1] = '1'; 561 ap[2] = '\0'; 562 break; 563 case '2': 564 if (ap[3] != '\0') 565 goto jbad; 566 ap[1] = '2'; 567 ap[2] = '\0'; 568 break; 569 case '\0': 570 break; 571 default: 572 jbad: errx(1, "illegal option -- %s", ap); 573 usage(); 574 } 575 break; 576 case 'o': 577 /* 578 * The original join allowed "-o arg arg". Convert to 579 * "-o arg -o arg". 580 */ 581 if (ap[2] != '\0') 582 break; 583 for (p = argv + 2; *p; ++p) { 584 if ((p[0][0] != '1' && p[0][0] != '2') || 585 p[0][1] != '.') 586 break; 587 len = strlen(*p); 588 if (len - 2 != strspn(*p + 2, "0123456789")) 589 break; 590 if ((t = malloc(len + 3)) == NULL) 591 enomem(); 592 t[0] = '-'; 593 t[1] = 'o'; 594 memmove(t + 2, *p, len + 1); 595 *p = t; 596 } 597 argv = p - 1; 598 break; 599 } 600 } 601 } 602 603 void 604 enomem() 605 { 606 errx(1, "no memory"); 607 } 608 609 void 610 usage() 611 { 612 (void)fprintf(stderr, "%s%s\n", 613 "usage: join [-a fileno | -v fileno ] [-e string] [-1 field] ", 614 "[-2 field]\n [-o list] [-t char] file1 file2"); 615 exit(1); 616 } 617
Indexes created Sun Sep 28 07:09:56 GMT 2025