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 * 3. 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: head/usr.bin/join/join.c 246319 2013-02-04 10:05:55Z andrew $
38 #include <sys/param.h>
51 * There's a structure per input file which encapsulates the state of the
52 * file. We repeatedly read lines from each file until we've read in all
53 * the consecutive lines from the file with a common join field. Then we
54 * compare the set of lines with an equivalent set from the other file.
57 char *line
; /* line */
58 u_long linealloc
; /* line allocated count */
59 char **fields
; /* line field(s) */
60 u_long fieldcnt
; /* line field(s) count */
61 u_long fieldalloc
; /* line field(s) allocated count */
65 FILE *fp
; /* file descriptor */
66 u_long joinf
; /* join field (-1, -2, -j) */
67 int unpair
; /* output unpairable lines (-a) */
68 u_long number
; /* 1 for file 1, 2 for file 2 */
70 LINE
*set
; /* set of lines with same field */
71 int pushbool
; /* if pushback is set */
72 u_long pushback
; /* line on the stack */
73 u_long setcnt
; /* set count */
74 u_long setalloc
; /* set allocated count */
76 static INPUT input1
= { NULL
, 0, 0, 1, NULL
, 0, 0, 0, 0 },
77 input2
= { NULL
, 0, 0, 2, NULL
, 0, 0, 0, 0 };
80 u_long filenum
; /* file number */
81 u_long fieldno
; /* field number */
83 static OLIST
*olist
; /* output field list */
84 static u_long olistcnt
; /* output field list count */
85 static u_long olistalloc
; /* output field allocated count */
87 static int joinout
= 1; /* show lines with matched join fields (-v) */
88 static int needsep
; /* need separator character */
89 static int spans
= 1; /* span multiple delimiters (-t) */
90 static char *empty
; /* empty field replacement string (-e) */
91 static wchar_t default_tabchar
[] = L
" \t";
92 static wchar_t *tabchar
= default_tabchar
; /* delimiter characters (-t) */
94 static int cmp(LINE
*, u_long
, LINE
*, u_long
);
95 static void fieldarg(char *);
96 static void joinlines(INPUT
*, INPUT
*);
97 static int mbscoll(const char *, const char *);
98 static char *mbssep(char **, const wchar_t *);
99 static void obsolete(char **);
100 static void outfield(LINE
*, u_long
, int);
101 static void outoneline(INPUT
*, LINE
*);
102 static void outtwoline(INPUT
*, LINE
*, INPUT
*, LINE
*);
103 static void slurp(INPUT
*);
104 static wchar_t *towcs(const char *);
105 static void usage(void);
108 main(int argc
, char *argv
[])
111 int aflag
, ch
, cval
, vflag
;
114 setlocale(LC_ALL
, "");
121 while ((ch
= getopt(argc
, argv
, "\01a:e:j:1:2:o:t:v:")) != -1) {
123 case '\01': /* See comment in obsolete(). */
125 F1
->unpair
= F2
->unpair
= 1;
128 if ((F1
->joinf
= strtol(optarg
, &end
, 10)) < 1)
129 errx(1, "-1 option field number less than 1");
131 errx(1, "illegal field number -- %s", optarg
);
135 if ((F2
->joinf
= strtol(optarg
, &end
, 10)) < 1)
136 errx(1, "-2 option field number less than 1");
138 errx(1, "illegal field number -- %s", optarg
);
143 switch(strtol(optarg
, &end
, 10)) {
151 errx(1, "-a option file number not 1 or 2");
155 errx(1, "illegal file number -- %s", optarg
);
161 if ((F1
->joinf
= F2
->joinf
=
162 strtol(optarg
, &end
, 10)) < 1)
163 errx(1, "-j option field number less than 1");
165 errx(1, "illegal field number -- %s", optarg
);
174 if (mbrtowc(&tabchar
[0], optarg
, MB_LEN_MAX
, NULL
) !=
176 errx(1, "illegal tab character specification");
182 switch (strtol(optarg
, &end
, 10)) {
190 errx(1, "-v option file number not 1 or 2");
194 errx(1, "illegal file number -- %s", optarg
);
205 errx(1, "the -a and -v options are mutually exclusive");
210 /* Open the files; "-" means stdin. */
211 if (!strcmp(*argv
, "-"))
213 else if ((F1
->fp
= fopen(*argv
, "r")) == NULL
)
216 if (!strcmp(*argv
, "-"))
218 else if ((F2
->fp
= fopen(*argv
, "r")) == NULL
)
220 if (F1
->fp
== stdin
&& F2
->fp
== stdin
)
221 errx(1, "only one input file may be stdin");
225 while (F1
->setcnt
&& F2
->setcnt
) {
226 cval
= cmp(F1
->set
, F1
->joinf
, F2
->set
, F2
->joinf
);
228 /* Oh joy, oh rapture, oh beauty divine! */
233 } else if (cval
< 0) {
234 /* File 1 takes the lead... */
239 /* File 2 takes the lead... */
247 * Now that one of the files is used up, optionally output any
248 * remaining lines from the other file.
266 LINE
*lp
, *lastlp
, tmp
;
272 * Read all of the lines from an input file that have the same
276 for (lastlp
= NULL
;; ++F
->setcnt
) {
278 * If we're out of space to hold line structures, allocate
279 * more. Initialize the structure so that we know that this
282 if (F
->setcnt
== F
->setalloc
) {
285 if ((F
->set
= realloc(F
->set
,
286 F
->setalloc
* sizeof(LINE
))) == NULL
)
288 memset(F
->set
+ cnt
, 0, 50 * sizeof(LINE
));
290 /* re-set lastlp in case it moved */
292 lastlp
= &F
->set
[F
->setcnt
- 1];
296 * Get any pushed back line, else get the next line. Allocate
297 * space as necessary. If taking the line from the stack swap
298 * the two structures so that we don't lose space allocated to
299 * either structure. This could be avoided by doing another
300 * level of indirection, but it's probably okay as is.
302 lp
= &F
->set
[F
->setcnt
];
304 lastlp
= &F
->set
[F
->setcnt
- 1];
306 tmp
= F
->set
[F
->setcnt
];
307 F
->set
[F
->setcnt
] = F
->set
[F
->pushback
];
308 F
->set
[F
->pushback
] = tmp
;
312 if ((bp
= fgetln(F
->fp
, &len
)) == NULL
)
314 if (lp
->linealloc
<= len
+ 1) {
315 lp
->linealloc
+= MAX(100, len
+ 1 - lp
->linealloc
);
317 realloc(lp
->line
, lp
->linealloc
)) == NULL
)
320 memmove(lp
->line
, bp
, len
);
322 /* Replace trailing newline, if it exists. */
323 if (bp
[len
- 1] == '\n')
324 lp
->line
[len
- 1] = '\0';
326 lp
->line
[len
] = '\0';
329 /* Split the line into fields, allocate space as necessary. */
331 while ((fieldp
= mbssep(&bp
, tabchar
)) != NULL
) {
332 if (spans
&& *fieldp
== '\0')
334 if (lp
->fieldcnt
== lp
->fieldalloc
) {
335 lp
->fieldalloc
+= 50;
336 if ((lp
->fields
= realloc(lp
->fields
,
337 lp
->fieldalloc
* sizeof(char *))) == NULL
)
340 lp
->fields
[lp
->fieldcnt
++] = fieldp
;
343 /* See if the join field value has changed. */
344 if (lastlp
!= NULL
&& cmp(lp
, F
->joinf
, lastlp
, F
->joinf
)) {
346 F
->pushback
= F
->setcnt
;
353 mbssep(char **stringp
, const wchar_t *delim
)
356 const wchar_t *spanp
;
360 if ((s
= *stringp
) == NULL
)
363 n
= mbrtowc(&c
, s
, MB_LEN_MAX
, NULL
);
364 if (n
== (size_t)-1 || n
== (size_t)-2)
365 errc(1, EILSEQ
, NULL
); /* XXX */
369 if ((sc
= *spanp
++) == c
) {
382 cmp(LINE
*lp1
, u_long fieldno1
, LINE
*lp2
, u_long fieldno2
)
384 if (lp1
->fieldcnt
<= fieldno1
)
385 return (lp2
->fieldcnt
<= fieldno2
? 0 : 1);
386 if (lp2
->fieldcnt
<= fieldno2
)
388 return (mbscoll(lp1
->fields
[fieldno1
], lp2
->fields
[fieldno2
]));
392 mbscoll(const char *s1
, const char *s2
)
398 return (strcoll(s1
, s2
));
399 if ((w1
= towcs(s1
)) == NULL
|| (w2
= towcs(s2
)) == NULL
)
400 err(1, NULL
); /* XXX */
401 ret
= wcscoll(w1
, w2
);
413 if ((n
= mbsrtowcs(NULL
, &s
, 0, NULL
)) == (size_t)-1)
415 if ((wcs
= malloc((n
+ 1) * sizeof(*wcs
))) == NULL
)
417 mbsrtowcs(wcs
, &s
, n
+ 1, NULL
);
422 joinlines(INPUT
*F1
, INPUT
*F2
)
427 * Output the results of a join comparison. The output may be from
428 * either file 1 or file 2 (in which case the first argument is the
429 * file from which to output) or from both.
432 for (cnt1
= 0; cnt1
< F1
->setcnt
; ++cnt1
)
433 outoneline(F1
, &F1
->set
[cnt1
]);
436 for (cnt1
= 0; cnt1
< F1
->setcnt
; ++cnt1
)
437 for (cnt2
= 0; cnt2
< F2
->setcnt
; ++cnt2
)
438 outtwoline(F1
, &F1
->set
[cnt1
], F2
, &F2
->set
[cnt2
]);
442 outoneline(INPUT
*F
, LINE
*lp
)
447 * Output a single line from one of the files, according to the
448 * join rules. This happens when we are writing unmatched single
449 * lines. Output empty fields in the right places.
452 for (cnt
= 0; cnt
< olistcnt
; ++cnt
) {
453 if (olist
[cnt
].filenum
== (unsigned)F
->number
)
454 outfield(lp
, olist
[cnt
].fieldno
, 0);
455 else if (olist
[cnt
].filenum
== 0)
456 outfield(lp
, F
->joinf
, 0);
461 for (cnt
= 0; cnt
< lp
->fieldcnt
; ++cnt
)
462 outfield(lp
, cnt
, 0);
470 outtwoline(INPUT
*F1
, LINE
*lp1
, INPUT
*F2
, LINE
*lp2
)
474 /* Output a pair of lines according to the join list (if any). */
476 for (cnt
= 0; cnt
< olistcnt
; ++cnt
)
477 if (olist
[cnt
].filenum
== 0) {
478 if (lp1
->fieldcnt
>= F1
->joinf
)
479 outfield(lp1
, F1
->joinf
, 0);
481 outfield(lp2
, F2
->joinf
, 0);
482 } else if (olist
[cnt
].filenum
== 1)
483 outfield(lp1
, olist
[cnt
].fieldno
, 0);
484 else /* if (olist[cnt].filenum == 2) */
485 outfield(lp2
, olist
[cnt
].fieldno
, 0);
488 * Output the join field, then the remaining fields from F1
491 outfield(lp1
, F1
->joinf
, 0);
492 for (cnt
= 0; cnt
< lp1
->fieldcnt
; ++cnt
)
493 if (F1
->joinf
!= cnt
)
494 outfield(lp1
, cnt
, 0);
495 for (cnt
= 0; cnt
< lp2
->fieldcnt
; ++cnt
)
496 if (F2
->joinf
!= cnt
)
497 outfield(lp2
, cnt
, 0);
506 outfield(LINE
*lp
, u_long fieldno
, int out_empty
)
509 printf("%lc", (wint_t)*tabchar
);
510 if (!ferror(stdout
)) {
511 if (lp
->fieldcnt
<= fieldno
|| out_empty
) {
515 if (*lp
->fields
[fieldno
] == '\0')
517 printf("%s", lp
->fields
[fieldno
]);
525 * Convert an output list argument "2.1, 1.3, 2.4" into an array of output
529 fieldarg(char *option
)
531 u_long fieldno
, filenum
;
534 while ((token
= strsep(&option
, ", \t")) != NULL
) {
538 filenum
= fieldno
= 0;
539 else if ((token
[0] == '1' || token
[0] == '2') &&
541 filenum
= token
[0] - '0';
542 fieldno
= strtol(token
+ 2, &end
, 10);
544 errx(1, "malformed -o option field");
546 errx(1, "field numbers are 1 based");
549 errx(1, "malformed -o option field");
550 if (olistcnt
== olistalloc
) {
552 if ((olist
= realloc(olist
,
553 olistalloc
* sizeof(OLIST
))) == NULL
)
556 olist
[olistcnt
].filenum
= filenum
;
557 olist
[olistcnt
].fieldno
= fieldno
;
563 obsolete(char **argv
)
568 while ((ap
= *++argv
) != NULL
) {
569 /* Return if "--". */
570 if (ap
[0] == '-' && ap
[1] == '-')
572 /* skip if not an option */
578 * The original join allowed "-a", which meant the
579 * same as -a1 plus -a2. POSIX 1003.2, Draft 11.2
580 * only specifies this as "-a 1" and "a -2", so we
581 * have to use another option flag, one that is
582 * unlikely to ever be used or accidentally entered
583 * on the command line. (Well, we could reallocate
584 * the argv array, but that hardly seems worthwhile.)
586 if (ap
[2] == '\0' && (argv
[1] == NULL
||
587 (strcmp(argv
[1], "1") != 0 &&
588 strcmp(argv
[1], "2") != 0))) {
590 warnx("-a option used without an argument; "
591 "reverting to historical behavior");
596 * The original join allowed "-j[12] arg" and "-j arg".
597 * Convert the former to "-[12] arg". Don't convert
598 * the latter since getopt(3) can handle it.
616 jbad
: errx(1, "illegal option -- %s", ap
);
622 * The original join allowed "-o arg arg".
623 * Convert to "-o arg -o arg".
627 for (p
= argv
+ 2; *p
; ++p
) {
628 if (p
[0][0] == '0' || ((p
[0][0] != '1' &&
629 p
[0][0] != '2') || p
[0][1] != '.'))
632 if (len
- 2 != strspn(*p
+ 2, "0123456789"))
634 if ((t
= malloc(len
+ 3)) == NULL
)
638 memmove(t
+ 2, *p
, len
+ 1);
650 fprintf(stderr
, "%s %s\n%s\n",
651 "usage: join [-a fileno | -v fileno ] [-e string] [-1 field]",
653 " [-o list] [-t char] file1 file2");