1 /* join - join lines of two files on a common field
2 Copyright (C) 91, 1995-2003 Free Software Foundation, Inc.
4 This program is free software; you can redistribute it and/or modify
5 it under the terms of the GNU General Public License as published by
6 the Free Software Foundation; either version 2, or (at your option)
9 This program is distributed in the hope that it will be useful,
10 but WITHOUT ANY WARRANTY; without even the implied warranty of
11 MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
12 GNU General Public License for more details.
14 You should have received a copy of the GNU General Public License
15 along with this program; if not, write to the Free Software Foundation,
16 Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA.
18 Written by Mike Haertel, mike@gnu.ai.mit.edu. */
24 #include <sys/types.h>
30 #include "hard-locale.h"
31 #include "linebuffer.h"
32 #include "memcasecmp.h"
36 /* The official name of this program (e.g., no `g' prefix). */
37 #define PROGRAM_NAME "join"
39 #define AUTHORS "Mike Haertel"
41 #define join system_join
43 /* Undefine, to avoid warning about redefinition on some systems. */
46 #define min(A, B) ((A) < (B) ? (A) : (B))
47 #define max(A, B) ((A) > (B) ? (A) : (B))
49 /* An element of the list identifying which fields to print for each
53 /* File number: 0, 1, or 2. 0 means use the join field.
54 1 means use the first file argument, 2 the second. */
57 /* Field index (zero-based), specified only when FILE is 1 or 2. */
63 /* A field of a line. */
66 const unsigned char *beg; /* First character in field. */
67 size_t len; /* The length of the field. */
70 /* A line read from an input file. */
73 struct linebuffer buf; /* The line itself. */
74 int nfields; /* Number of elements in `fields'. */
75 int nfields_allocated; /* Number of elements in `fields'. */
79 /* One or more consecutive lines read from a file that all have the
80 same join field value. */
83 int count; /* Elements used in `lines'. */
84 int alloc; /* Elements allocated in `lines'. */
88 /* The name this program was run with. */
91 /* Nonzero if the LC_COLLATE locale is hard. */
92 static int hard_LC_COLLATE;
94 /* If nonzero, print unpairable lines in file 1 or 2. */
95 static int print_unpairables_1, print_unpairables_2;
97 /* If nonzero, print pairable lines. */
98 static int print_pairables;
100 /* Empty output field filler. */
101 static char *empty_filler;
103 /* Field to join on. */
104 static int join_field_1, join_field_2;
106 /* List of fields to print. */
107 static struct outlist outlist_head;
109 /* Last element in `outlist', where a new element can be added. */
110 static struct outlist *outlist_end = &outlist_head;
112 /* Tab character separating fields; if this is NUL fields are separated
113 by any nonempty string of white space, otherwise by exactly one
115 static unsigned char tab;
117 /* When using getopt_long_only, no long option can start with
118 a character that is a short option. */
119 static struct option const longopts[] =
121 {"ignore-case", no_argument, NULL, 'i'},
122 {"j", required_argument, NULL, 'j'},
123 {"j1", required_argument, NULL, '1'},
124 {"j2", required_argument, NULL, '2'},
125 {GETOPT_HELP_OPTION_DECL},
126 {GETOPT_VERSION_OPTION_DECL},
130 /* Used to print non-joining lines */
131 static struct line uni_blank;
133 /* If nonzero, ignore case when comparing join fields. */
134 static int ignore_case;
140 fprintf (stderr, _("Try `%s --help' for more information.\n"),
145 Usage: %s [OPTION]... FILE1 FILE2\n\
149 For each pair of input lines with identical join fields, write a line to\n\
150 standard output. The default join field is the first, delimited\n\
151 by whitespace. When FILE1 or FILE2 (not both) is -, read standard input.\n\
153 -a FILENUM print unpairable lines coming from file FILENUM, where\n\
154 FILENUM is 1 or 2, corresponding to FILE1 or FILE2\n\
155 -e EMPTY replace missing input fields with EMPTY\n\
158 -i, --ignore-case ignore differences in case when comparing fields\n\
159 -j FIELD (obsolescent) equivalent to `-1 FIELD -2 FIELD'\n\
160 -j1 FIELD (obsolescent) equivalent to `-1 FIELD'\n\
161 -j2 FIELD (obsolescent) equivalent to `-2 FIELD'\n\
162 -o FORMAT obey FORMAT while constructing output line\n\
163 -t CHAR use CHAR as input and output field separator\n\
166 -v FILENUM like -a FILENUM, but suppress joined output lines\n\
167 -1 FIELD join on this FIELD of file 1\n\
168 -2 FIELD join on this FIELD of file 2\n\
170 fputs (HELP_OPTION_DESCRIPTION, stdout);
171 fputs (VERSION_OPTION_DESCRIPTION, stdout);
174 Unless -t CHAR is given, leading blanks separate fields and are ignored,\n\
175 else fields are separated by CHAR. Any FIELD is a field number counted\n\
176 from 1. FORMAT is one or more comma or blank separated specifications,\n\
177 each being `FILENUM.FIELD' or `0'. Default FORMAT outputs the join field,\n\
178 the remaining fields from FILE1, the remaining fields from FILE2, all\n\
179 separated by CHAR.\n\
181 printf (_("\nReport bugs to <%s>.\n"), PACKAGE_BUGREPORT);
183 exit (status == 0 ? EXIT_SUCCESS : EXIT_FAILURE);
187 ADD_FIELD (struct line *line, const unsigned char *field, size_t len)
189 if (line->nfields >= line->nfields_allocated)
191 line->nfields_allocated = (3 * line->nfields_allocated) / 2 + 1;
192 line->fields = xrealloc (line->fields,
193 (line->nfields_allocated
194 * sizeof (struct field)));
196 line->fields[line->nfields].beg = field;
197 line->fields[line->nfields].len = len;
201 /* Fill in the `fields' structure in LINE. */
204 xfields (struct line *line)
207 unsigned char *ptr0 = (unsigned char *) line->buf.buffer;
212 lim = ptr0 + line->buf.length - 1;
216 /* Skip leading blanks before the first field. */
217 while (ptr < lim && ISBLANK (*ptr))
221 for (i = 0; ptr < lim; ++i)
228 while (ptr < lim && *ptr != tab)
230 ADD_FIELD (line, beg, ptr - beg);
239 while (ptr < lim && !ISBLANK (*ptr))
241 ADD_FIELD (line, beg, ptr - beg);
242 while (ptr < lim && ISBLANK (*ptr))
247 if (ptr != ptr0 && ((!tab && ISBLANK (ptr[-1])) || ptr[-1] == tab))
249 /* Add one more (empty) field because the last character of the
250 line was a delimiter. */
251 ADD_FIELD (line, NULL, 0);
255 /* Read a line from FP into LINE and split it into fields.
256 Return 0 if EOF, 1 otherwise. */
259 get_line (FILE *fp, struct line *line)
261 initbuffer (&line->buf);
263 if (! readline (&line->buf, fp))
265 free (line->buf.buffer);
266 line->buf.buffer = NULL;
270 line->nfields_allocated = 0;
278 freeline (struct line *line)
281 free (line->buf.buffer);
282 line->buf.buffer = NULL;
286 initseq (struct seq *seq)
290 seq->lines = xmalloc (seq->alloc * sizeof (struct line));
293 /* Read a line from FP and add it to SEQ. Return 0 if EOF, 1 otherwise. */
296 getseq (FILE *fp, struct seq *seq)
298 if (seq->count == seq->alloc)
301 seq->lines = (struct line *)
302 xrealloc (seq->lines, seq->alloc * sizeof (struct line));
305 if (get_line (fp, &seq->lines[seq->count]))
314 delseq (struct seq *seq)
317 for (i = 0; i < seq->count; i++)
318 if (seq->lines[i].buf.buffer)
319 freeline (&seq->lines[i]);
323 /* Return <0 if the join field in LINE1 compares less than the one in LINE2;
324 >0 if it compares greater; 0 if it compares equal.
325 Report an error and exit if the comparison fails. */
328 keycmp (struct line *line1, struct line *line2)
330 /* Start of field to compare in each file. */
331 const unsigned char *beg1, *beg2;
333 size_t len1, len2; /* Length of fields to compare. */
336 if (join_field_1 < line1->nfields)
338 beg1 = line1->fields[join_field_1].beg;
339 len1 = line1->fields[join_field_1].len;
347 if (join_field_2 < line2->nfields)
349 beg2 = line2->fields[join_field_2].beg;
350 len2 = line2->fields[join_field_2].len;
359 return len2 == 0 ? 0 : -1;
363 /* Use an if-statement here rather than a function variable to
364 avoid portability hassles of getting a non-conflicting declaration
368 /* FIXME: ignore_case does not work with NLS (in particular,
369 with multibyte chars). */
370 diff = memcasecmp (beg1, beg2, min (len1, len2));
374 if (HAVE_SETLOCALE && hard_LC_COLLATE)
375 return xmemcoll ((char *) beg1, len1, (char *) beg2, len2);
376 diff = memcmp (beg1, beg2, min (len1, len2));
381 return len1 < len2 ? -1 : len1 != len2;
384 /* Print field N of LINE if it exists and is nonempty, otherwise
385 `empty_filler' if it is nonempty. */
388 prfield (int n, struct line *line)
392 if (n < line->nfields)
394 len = line->fields[n].len;
396 fwrite (line->fields[n].beg, 1, len, stdout);
397 else if (empty_filler)
398 fputs (empty_filler, stdout);
400 else if (empty_filler)
401 fputs (empty_filler, stdout);
404 /* Print the join of LINE1 and LINE2. */
407 prjoin (struct line *line1, struct line *line2)
409 const struct outlist *outlist;
411 outlist = outlist_head.next;
414 const struct outlist *o;
424 if (line1 == &uni_blank)
427 field = join_field_2;
432 field = join_field_1;
437 line = (o->file == 1 ? line1 : line2);
438 assert (o->field >= 0);
441 prfield (field, line);
445 putchar (tab ? tab : ' ');
453 if (line1 == &uni_blank)
460 prfield (join_field_1, line1);
461 for (i = 0; i < join_field_1 && i < line1->nfields; ++i)
463 putchar (tab ? tab : ' ');
466 for (i = join_field_1 + 1; i < line1->nfields; ++i)
468 putchar (tab ? tab : ' ');
472 for (i = 0; i < join_field_2 && i < line2->nfields; ++i)
474 putchar (tab ? tab : ' ');
477 for (i = join_field_2 + 1; i < line2->nfields; ++i)
479 putchar (tab ? tab : ' ');
486 /* Print the join of the files in FP1 and FP2. */
489 join (FILE *fp1, FILE *fp2)
491 struct seq seq1, seq2;
493 int diff, i, j, eof1, eof2;
495 /* Read the first line of each file. */
501 while (seq1.count && seq2.count)
503 diff = keycmp (&seq1.lines[0], &seq2.lines[0]);
506 if (print_unpairables_1)
507 prjoin (&seq1.lines[0], &uni_blank);
508 freeline (&seq1.lines[0]);
515 if (print_unpairables_2)
516 prjoin (&uni_blank, &seq2.lines[0]);
517 freeline (&seq2.lines[0]);
523 /* Keep reading lines from file1 as long as they continue to
524 match the current line from file2. */
527 if (!getseq (fp1, &seq1))
533 while (!keycmp (&seq1.lines[seq1.count - 1], &seq2.lines[0]));
535 /* Keep reading lines from file2 as long as they continue to
536 match the current line from file1. */
539 if (!getseq (fp2, &seq2))
545 while (!keycmp (&seq1.lines[0], &seq2.lines[seq2.count - 1]));
549 for (i = 0; i < seq1.count - 1; ++i)
550 for (j = 0; j < seq2.count - 1; ++j)
551 prjoin (&seq1.lines[i], &seq2.lines[j]);
554 for (i = 0; i < seq1.count - 1; ++i)
555 freeline (&seq1.lines[i]);
558 seq1.lines[0] = seq1.lines[seq1.count - 1];
564 for (i = 0; i < seq2.count - 1; ++i)
565 freeline (&seq2.lines[i]);
568 seq2.lines[0] = seq2.lines[seq2.count - 1];
575 if (print_unpairables_1 && seq1.count)
577 prjoin (&seq1.lines[0], &uni_blank);
578 freeline (&seq1.lines[0]);
579 while (get_line (fp1, &line))
581 prjoin (&line, &uni_blank);
586 if (print_unpairables_2 && seq2.count)
588 prjoin (&uni_blank, &seq2.lines[0]);
589 freeline (&seq2.lines[0]);
590 while (get_line (fp2, &line))
592 prjoin (&uni_blank, &line);
601 /* Add a field spec for field FIELD of file FILE to `outlist'. */
604 add_field (int file, int field)
608 assert (file == 0 || file == 1 || file == 2);
609 assert (file == 0 ? field < 0 : field >= 0);
611 o = xmalloc (sizeof (struct outlist));
616 /* Add to the end of the list so the fields are in the right order. */
617 outlist_end->next = o;
621 /* Convert a single field specifier string, S, to a *FILE_INDEX, *FIELD_INDEX
622 pair. In S, the field index string is 1-based; *FIELD_INDEX is zero-based.
623 If S is valid, return zero. Otherwise, give a diagnostic, don't update
624 *FILE_INDEX or *FIELD_INDEX, and return nonzero. */
627 decode_field_spec (const char *s, int *file_index, int *field_index)
631 /* The first character must be 0, 1, or 2. */
638 /* Give *field_index an invalid value. */
644 /* `0' must be all alone -- no `.FIELD'. */
645 error (0, 0, _("invalid field specifier: `%s'"), s);
651 if (s[1] == '.' && s[2] != '\0')
656 s_err = xstrtol (s + 2, NULL, 10, &tmp_long, "");
657 if (s_err != LONGINT_OK || tmp_long <= 0 || tmp_long > INT_MAX)
659 error (0, 0, _("invalid field number: `%s'"), s + 2);
663 *file_index = s[0] - '0';
664 /* Convert to a zero-based index. */
665 *field_index = (int) tmp_long - 1;
672 error (0, 0, _("invalid file number in field spec: `%s'"), s);
678 /* Add the comma or blank separated field spec(s) in STR to `outlist'.
679 Return nonzero to indicate failure. */
682 add_field_list (const char *c_str)
686 /* Make a writable copy of c_str. */
687 str = (char *) alloca (strlen (c_str) + 1);
694 int file_index, field_index;
697 p = strpbrk (p, ", \t");
700 invalid = decode_field_spec (spec_item, &file_index, &field_index);
703 add_field (file_index, field_index);
704 uni_blank.nfields = max (uni_blank.nfields, field_index);
710 /* Create a blank line with COUNT fields separated by tabs. */
713 make_blank (struct line *blank, int count)
716 unsigned char *buffer;
717 struct field *fields;
718 blank->nfields = count;
719 blank->buf.size = blank->buf.length = count + 1;
720 blank->buf.buffer = xmalloc (blank->buf.size);
721 buffer = (unsigned char *) blank->buf.buffer;
722 blank->fields = fields =
723 xmalloc (sizeof (struct field) * count);
724 for (i = 0; i < count; i++)
727 fields[i].beg = &buffer[i];
734 main (int argc, char **argv)
738 int optc, prev_optc = 0, nfiles;
740 initialize_main (&argc, &argv);
741 program_name = argv[0];
742 setlocale (LC_ALL, "");
743 bindtextdomain (PACKAGE, LOCALEDIR);
744 textdomain (PACKAGE);
745 hard_LC_COLLATE = hard_locale (LC_COLLATE);
747 atexit (close_stdout);
749 /* Initialize this before parsing options. In parsing options,
750 it may be increased. */
751 uni_blank.nfields = 1;
756 while ((optc = getopt_long_only (argc, argv, "-a:e:i1:2:o:t:v:", longopts,
771 if (xstrtol (optarg, NULL, 10, &val, "") != LONGINT_OK
772 || (val != 1 && val != 2))
773 error (EXIT_FAILURE, 0, _("invalid field number: `%s'"), optarg);
775 print_unpairables_1 = 1;
777 print_unpairables_2 = 1;
781 empty_filler = optarg;
789 if (xstrtol (optarg, NULL, 10, &val, "") != LONGINT_OK
790 || val <= 0 || val > INT_MAX)
792 error (EXIT_FAILURE, 0,
793 _("invalid field number for file 1: `%s'"), optarg);
795 join_field_1 = (int) val - 1;
799 if (xstrtol (optarg, NULL, 10, &val, "") != LONGINT_OK
800 || val <= 0 || val > INT_MAX)
801 error (EXIT_FAILURE, 0,
802 _("invalid field number for file 2: `%s'"), optarg);
803 join_field_2 = (int) val - 1;
807 if (xstrtol (optarg, NULL, 10, &val, "") != LONGINT_OK
808 || val <= 0 || val > INT_MAX)
809 error (EXIT_FAILURE, 0, _("invalid field number: `%s'"), optarg);
810 join_field_1 = join_field_2 = (int) val - 1;
814 if (add_field_list (optarg))
822 case 1: /* Non-option argument. */
823 if (prev_optc == 'o' && optind <= argc - 2)
825 if (add_field_list (optarg))
828 /* Might be continuation of args to -o. */
829 continue; /* Don't change `prev_optc'. */
834 error (0, 0, _("too many non-option arguments"));
835 usage (EXIT_FAILURE);
837 names[nfiles++] = optarg;
840 case_GETOPT_HELP_CHAR;
842 case_GETOPT_VERSION_CHAR (PROGRAM_NAME, AUTHORS);
845 usage (EXIT_FAILURE);
850 /* Now that we've seen the options, we can construct the blank line
852 make_blank (&uni_blank, uni_blank.nfields);
856 error (0, 0, _("too few non-option arguments"));
857 usage (EXIT_FAILURE);
860 fp1 = STREQ (names[0], "-") ? stdin : fopen (names[0], "r");
862 error (EXIT_FAILURE, errno, "%s", names[0]);
863 fp2 = STREQ (names[1], "-") ? stdin : fopen (names[1], "r");
865 error (EXIT_FAILURE, errno, "%s", names[1]);
867 error (EXIT_FAILURE, errno, _("both files cannot be standard input"));
870 if (fp1 != stdin && fclose (fp1) == EOF)
871 error (EXIT_FAILURE, errno, "%s", names[0]);
872 if (fp2 != stdin && fclose (fp2) == EOF)
873 error (EXIT_FAILURE, errno, "%s", names[1]);
874 if ((fp1 == stdin || fp2 == stdin) && fclose (stdin) == EOF)
875 error (EXIT_FAILURE, errno, "-");