2 * Copyright (c) 1991, 1993, 1994
3 * The Regents of the University of California. All rights reserved.
5 * This code is derived from software contributed to Berkeley by
6 * Steve Hayman of the Computer Science Department, Indiana University,
7 * Michiro Hikida and David Goodenough.
9 * Redistribution and use in source and binary forms, with or without
10 * modification, are permitted provided that the following conditions
12 * 1. Redistributions of source code must retain the above copyright
13 * notice, this list of conditions and the following disclaimer.
14 * 2. Redistributions in binary form must reproduce the above copyright
15 * notice, this list of conditions and the following disclaimer in the
16 * documentation and/or other materials provided with the distribution.
17 * 4. Neither the name of the University nor the names of its contributors
18 * may be used to endorse or promote products derived from this software
19 * without specific prior written permission.
21 * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
22 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
23 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
24 * ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
25 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
26 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
27 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
28 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
29 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
30 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
33 * @(#) Copyright (c) 1991, 1993, 1994 The Regents of the University of California. All rights reserved.
34 * @(#)join.c 8.6 (Berkeley) 5/4/95
35 * $FreeBSD: src/usr.bin/join/join.c,v 1.10.2.1 2002/06/18 05:14:49 jmallett Exp $
36 * $DragonFly: src/usr.bin/join/join.c,v 1.6 2005/01/16 21:40:42 cpressey Exp $
39 #include <sys/param.h>
50 * There's a structure per input file which encapsulates the state of the
51 * file. We repeatedly read lines from each file until we've read in all
52 * the consecutive lines from the file with a common join field. Then we
53 * compare the set of lines with an equivalent set from the other file.
56 char *line; /* line */
57 u_long linealloc; /* line allocated count */
58 char **fields; /* line field(s) */
59 u_long fieldcnt; /* line field(s) count */
60 u_long fieldalloc; /* line field(s) allocated count */
64 FILE *fp; /* file descriptor */
65 u_long joinf; /* join field (-1, -2, -j) */
66 int unpair; /* output unpairable lines (-a) */
67 int number; /* 1 for file 1, 2 for file 2 */
69 LINE *set; /* set of lines with same field */
70 int pushbool; /* if pushback is set */
71 u_long pushback; /* line on the stack */
72 u_long setcnt; /* set count */
73 u_long setalloc; /* set allocated count */
75 INPUT input1 = { NULL, 0, 0, 1, NULL, 0, 0, 0, 0 },
76 input2 = { NULL, 0, 0, 2, NULL, 0, 0, 0, 0 };
79 u_long filenum; /* file number */
80 u_long fieldno; /* field number */
82 OLIST *olist; /* output field list */
83 u_long olistcnt; /* output field list count */
84 u_long olistalloc; /* output field allocated count */
86 int joinout = 1; /* show lines with matched join fields (-v) */
87 int needsep; /* need separator character */
88 int spans = 1; /* span multiple delimiters (-t) */
89 char *empty; /* empty field replacement string (-e) */
90 static char default_tabchar[] = " \t";
91 char *tabchar = default_tabchar;/* delimiter characters (-t) */
93 int cmp(LINE *, u_long, LINE *, u_long);
94 void fieldarg(char *);
95 void joinlines(INPUT *, INPUT *);
96 void obsolete(char **);
97 void outfield(LINE *, u_long, int);
98 void outoneline(INPUT *, LINE *);
99 void outtwoline(INPUT *, LINE *, INPUT *, LINE *);
104 main(int argc, char **argv)
107 int aflag, ch, cval, vflag;
110 setlocale(LC_ALL, "");
117 while ((ch = getopt(argc, argv, "\01a:e:j:1:2:o:t:v:")) != -1) {
119 case '\01': /* See comment in obsolete(). */
121 F1->unpair = F2->unpair = 1;
124 if ((F1->joinf = strtol(optarg, &end, 10)) < 1)
125 errx(1, "-1 option field number less than 1");
127 errx(1, "illegal field number -- %s", optarg);
131 if ((F2->joinf = strtol(optarg, &end, 10)) < 1)
132 errx(1, "-2 option field number less than 1");
134 errx(1, "illegal field number -- %s", optarg);
139 switch(strtol(optarg, &end, 10)) {
147 errx(1, "-a option file number not 1 or 2");
151 errx(1, "illegal file number -- %s", optarg);
157 if ((F1->joinf = F2->joinf =
158 strtol(optarg, &end, 10)) < 1)
159 errx(1, "-j option field number less than 1");
161 errx(1, "illegal field number -- %s", optarg);
170 if (strlen(tabchar = optarg) != 1)
171 errx(1, "illegal tab character specification");
176 switch (strtol(optarg, &end, 10)) {
184 errx(1, "-v option file number not 1 or 2");
188 errx(1, "illegal file number -- %s", optarg);
199 errx(1, "the -a and -v options are mutually exclusive");
204 /* Open the files; "-" means stdin. */
205 if (!strcmp(*argv, "-"))
207 else if ((F1->fp = fopen(*argv, "r")) == NULL)
210 if (!strcmp(*argv, "-"))
212 else if ((F2->fp = fopen(*argv, "r")) == NULL)
214 if (F1->fp == stdin && F2->fp == stdin)
215 errx(1, "only one input file may be stdin");
219 while (F1->setcnt && F2->setcnt) {
220 cval = cmp(F1->set, F1->joinf, F2->set, F2->joinf);
222 /* Oh joy, oh rapture, oh beauty divine! */
227 } else if (cval < 0) {
228 /* File 1 takes the lead... */
233 /* File 2 takes the lead... */
241 * Now that one of the files is used up, optionally output any
242 * remaining lines from the other file.
260 LINE *lp, *lastlp, tmp;
266 * Read all of the lines from an input file that have the same
270 for (lastlp = NULL;; ++F->setcnt) {
272 * If we're out of space to hold line structures, allocate
273 * more. Initialize the structure so that we know that this
276 if (F->setcnt == F->setalloc) {
279 if ((F->set = realloc(F->set,
280 F->setalloc * sizeof(LINE))) == NULL)
282 memset(F->set + cnt, 0, 50 * sizeof(LINE));
284 /* re-set lastlp in case it moved */
286 lastlp = &F->set[F->setcnt - 1];
290 * Get any pushed back line, else get the next line. Allocate
291 * space as necessary. If taking the line from the stack swap
292 * the two structures so that we don't lose space allocated to
293 * either structure. This could be avoided by doing another
294 * level of indirection, but it's probably okay as is.
296 lp = &F->set[F->setcnt];
298 lastlp = &F->set[F->setcnt - 1];
300 tmp = F->set[F->setcnt];
301 F->set[F->setcnt] = F->set[F->pushback];
302 F->set[F->pushback] = tmp;
306 if ((bp = fgetln(F->fp, &len)) == NULL)
308 if (lp->linealloc <= len + 1) {
309 lp->linealloc += MAX(100, len + 1 - lp->linealloc);
311 realloc(lp->line, lp->linealloc)) == NULL)
314 memmove(lp->line, bp, len);
316 /* Replace trailing newline, if it exists. */
317 if (bp[len - 1] == '\n')
318 lp->line[len - 1] = '\0';
320 lp->line[len] = '\0';
323 /* Split the line into fields, allocate space as necessary. */
325 while ((fieldp = strsep(&bp, tabchar)) != NULL) {
326 if (spans && *fieldp == '\0')
328 if (lp->fieldcnt == lp->fieldalloc) {
329 lp->fieldalloc += 50;
330 if ((lp->fields = realloc(lp->fields,
331 lp->fieldalloc * sizeof(char *))) == NULL)
334 lp->fields[lp->fieldcnt++] = fieldp;
337 /* See if the join field value has changed. */
338 if (lastlp != NULL && cmp(lp, F->joinf, lastlp, F->joinf)) {
340 F->pushback = F->setcnt;
347 cmp(LINE *lp1, u_long fieldno1, LINE *lp2, u_long fieldno2)
349 if (lp1->fieldcnt <= fieldno1)
350 return (lp2->fieldcnt <= fieldno2 ? 0 : 1);
351 if (lp2->fieldcnt <= fieldno2)
353 return (strcoll(lp1->fields[fieldno1], lp2->fields[fieldno2]));
357 joinlines(INPUT *F1, INPUT *F2)
359 unsigned int cnt1, cnt2;
362 * Output the results of a join comparison. The output may be from
363 * either file 1 or file 2 (in which case the first argument is the
364 * file from which to output) or from both.
367 for (cnt1 = 0; cnt1 < F1->setcnt; ++cnt1)
368 outoneline(F1, &F1->set[cnt1]);
371 for (cnt1 = 0; cnt1 < F1->setcnt; ++cnt1)
372 for (cnt2 = 0; cnt2 < F2->setcnt; ++cnt2)
373 outtwoline(F1, &F1->set[cnt1], F2, &F2->set[cnt2]);
377 outoneline(INPUT *F, LINE *lp)
382 * Output a single line from one of the files, according to the
383 * join rules. This happens when we are writing unmatched single
384 * lines. Output empty fields in the right places.
387 for (cnt = 0; cnt < olistcnt; ++cnt) {
388 if (olist[cnt].filenum == (unsigned)F->number)
389 outfield(lp, olist[cnt].fieldno, 0);
390 else if (olist[cnt].filenum == 0)
391 outfield(lp, F->joinf, 0);
396 for (cnt = 0; cnt < lp->fieldcnt; ++cnt)
397 outfield(lp, cnt, 0);
405 outtwoline(INPUT *F1, LINE *lp1, INPUT *F2, LINE *lp2)
409 /* Output a pair of lines according to the join list (if any). */
411 for (cnt = 0; cnt < olistcnt; ++cnt)
412 if (olist[cnt].filenum == 0) {
413 if (lp1->fieldcnt >= F1->joinf)
414 outfield(lp1, F1->joinf, 0);
416 outfield(lp2, F2->joinf, 0);
417 } else if (olist[cnt].filenum == 1)
418 outfield(lp1, olist[cnt].fieldno, 0);
419 else /* if (olist[cnt].filenum == 2) */
420 outfield(lp2, olist[cnt].fieldno, 0);
423 * Output the join field, then the remaining fields from F1
426 outfield(lp1, F1->joinf, 0);
427 for (cnt = 0; cnt < lp1->fieldcnt; ++cnt)
428 if (F1->joinf != cnt)
429 outfield(lp1, cnt, 0);
430 for (cnt = 0; cnt < lp2->fieldcnt; ++cnt)
431 if (F2->joinf != cnt)
432 outfield(lp2, cnt, 0);
441 outfield(LINE *lp, u_long fieldno, int out_empty)
444 (void)printf("%c", *tabchar);
445 if (!ferror(stdout)) {
446 if (lp->fieldcnt <= fieldno || out_empty) {
448 (void)printf("%s", empty);
450 if (*lp->fields[fieldno] == '\0')
452 (void)printf("%s", lp->fields[fieldno]);
460 * Convert an output list argument "2.1, 1.3, 2.4" into an array of output
464 fieldarg(char *option)
466 u_long fieldno, filenum;
469 while ((token = strsep(&option, ", \t")) != NULL) {
473 filenum = fieldno = 0;
474 else if ((token[0] == '1' || token[0] == '2') &&
476 filenum = token[0] - '0';
477 fieldno = strtol(token + 2, &end, 10);
479 errx(1, "malformed -o option field");
481 errx(1, "field numbers are 1 based");
484 errx(1, "malformed -o option field");
485 if (olistcnt == olistalloc) {
487 if ((olist = realloc(olist,
488 olistalloc * sizeof(OLIST))) == NULL)
491 olist[olistcnt].filenum = filenum;
492 olist[olistcnt].fieldno = fieldno;
498 obsolete(char **argv)
503 while ((ap = *++argv) != NULL) {
504 /* Return if "--". */
505 if (ap[0] == '-' && ap[1] == '-')
507 /* skip if not an option */
513 * The original join allowed "-a", which meant the
514 * same as -a1 plus -a2. POSIX 1003.2, Draft 11.2
515 * only specifies this as "-a 1" and "a -2", so we
516 * have to use another option flag, one that is
517 * unlikely to ever be used or accidentally entered
518 * on the command line. (Well, we could reallocate
519 * the argv array, but that hardly seems worthwhile.)
521 if (ap[2] == '\0' && (argv[1] == NULL ||
522 (strcmp(argv[1], "1") != 0 &&
523 strcmp(argv[1], "2") != 0))) {
525 warnx("-a option used without an argument; "
526 "reverting to historical behavior");
531 * The original join allowed "-j[12] arg" and "-j arg".
532 * Convert the former to "-[12] arg". Don't convert
533 * the latter since getopt(3) can handle it.
551 jbad: errx(1, "illegal option -- %s", ap);
557 * The original join allowed "-o arg arg".
558 * Convert to "-o arg -o arg".
562 for (p = argv + 2; *p; ++p) {
563 if ((p[0][0] != '1' && p[0][0] != '2') ||
567 if (len - 2 != strspn(*p + 2, "0123456789"))
569 if ((t = malloc(len + 3)) == NULL)
573 memmove(t + 2, *p, len + 1);
585 fprintf(stderr, "%s",
586 "usage: join [-a fileno | -v fileno ] [-e string] [-j fileno field]\n"
587 " [-1 field] [-2 field] [-o list] [-t char] file1 file2\n");