Replacing join with NetBSD join

Change-Id: I1257736f755eab0ce39124f0d78bfa48426ba820
This commit is contained in:
Lionel Sambuc 2012-12-04 15:07:20 +01:00
parent c75851fccb
commit 6e0ed9c90c
12 changed files with 864 additions and 487 deletions

View file

@ -12,7 +12,7 @@ SUBDIR= add_route arp ash at backup banner basename btrace cal \
find finger fingerd fix fold format fortune fsck.mfs \
ftp101 gcore gcov-pull getty grep head hexdump host \
hostaddr id ifconfig ifdef \
intr ipcrm ipcs irdpd isoread join kill last \
intr ipcrm ipcs irdpd isoread kill last \
less loadkeys loadramdisk logger look lp \
lpd ls lspci mail MAKEDEV \
mesg mined mkfifo \

View file

@ -1,4 +0,0 @@
PROG= join
MAN=
.include <bsd.prog.mk>

View file

@ -1,360 +0,0 @@
/* join - relation data base operator Author: Saeko Hirabayashi */
/* Written by Saeko Hirabayashi, 1989.
* 1992-01-28 Modified by Kouichi Hirabayashi to add some POSIX1003.2 options.
*
* This a free program.
*/
#include <string.h>
#include <stdio.h>
#define MAXFLD 200 /* maximum # of fields to accept */
int main(int argc, char **argv);
void error(char *s, char *t);
void usage(void);
void match(void);
void f1_only(void);
void f2_only(void);
void output(int flag);
void outfld(int file);
void outputf(int flag);
int compare(void);
int get1(void);
int get2(int back);
int getrec(int file);
int split(int file);
int atoi(char *str);
int exit(int val);
FILE * efopen(char *file, char *mode);
void(*outfun) (int file);
#define F1 1
#define F2 2
#define SEP (sep ? sep : ' ')
FILE *fp[2]; /* file pointer for file1 and file2 */
long head; /* head of the current (same)key group of the
* file2 */
char buf[2][BUFSIZ]; /* input buffer for file1 and file2 */
char *fld[2][MAXFLD]; /* field vector for file1 and file2 */
int nfld[2]; /* # of fields for file1 and file2 */
int kpos[2]; /* key field position for file1 and file2
* (from 0) */
char oldkey[BUFSIZ]; /* previous key of the file1 */
struct { /* output list by -o option */
int o_file; /* file #: 0 or 1 */
int o_field; /* field #: 0, 1, 2, .. */
} olist[MAXFLD];
int nout; /* # of output filed */
int aflag; /* n for '-an': F1 or F2 or both */
int vflag; /* n for '-vn': F1 or F2 or both */
char *es; /* s for '-e s' */
char sep; /* c for -tc: filed separator */
char *cmd; /* name of this program */
int main(argc, argv)
int argc;
char **argv;
{
register char *s;
int c, i, j;
cmd = argv[0];
outfun = output; /* default output form */
while (--argc > 0 && (*++argv)[0] == '-' && (*argv)[1]) {
/* "-" is a file name (stdin) */
s = argv[0] + 1;
if ((c = *s) == '-' && !s[1]) {
++argv;
--argc;
break; /* -- */
}
if (*++s == '\0') {
s = *++argv;
--argc;
}
switch (c) {
case 'a': /* add unpairable line to output */
vflag = 0;
switch (*s) {
case '1': aflag |= F1; break;
case '2': aflag |= F2; break;
default: aflag |= (F1 | F2); break;
}
break;
case 'e': /* replace empty field by es */
es = s;
break;
case 'j': /* key field (obsolute) */
c = *s++;
if (*s == '\0') {
s = *++argv;
--argc;
}
case '1': /* key field of file1 */
case '2': /* key field of file2 */
i = atoi(s) - 1;
switch (c) {
case '1': kpos[0] = i; break;
case '2': kpos[1] = i; break;
default: kpos[0] = kpos[1] = i;
break;
}
break;
case 'o': /* specify output format */
do {
i = j = 0;
sscanf(s, "%d.%d", &i, &j);
if (i < 1 || j < 1 || i > 2) usage();
olist[nout].o_file = i - 1;
olist[nout].o_field = j - 1;
nout++;
if ((s = strchr(s, ',')) != (char *) 0)
s++;
else {
s = *++argv;
--argc;
}
} while (argc > 2 && *s != '-');
++argc;
--argv; /* compensation */
outfun = outputf;
break;
case 't': /* tab char */
sep = *s;
break;
case 'v': /* output unpairable line only */
aflag = 0;
switch (*s) {
case '1': vflag |= F1; break;
case '2': vflag |= F2; break;
default: vflag |= (F1 | F2); break;
}
break;
default: usage();
}
}
if (argc != 2) usage();
fp[0] = strcmp(argv[0], "-") ? efopen(argv[0], "r") : stdin;
fp[1] = efopen(argv[1], "r");
nfld[0] = get1(); /* read file1 */
nfld[1] = get2(0); /* read file2 */
while (nfld[0] || nfld[1]) {
if ((i = compare()) == 0)
match();
else if (i < 0)
f1_only();
else
f2_only();
}
fflush(stdout);
exit(0);
}
void usage()
{
fprintf(stderr,
"Usage: %s [-an|-vn] [-e str] [-o list] [-tc] [-1 f] [-2 f] file1 file2\n",
cmd);
exit(1);
}
int compare()
{ /* compare key field */
register int r;
if (nfld[1] == 0) /* file2 EOF */
r = -1;
else if (nfld[0] == 0) /* file1 EOF */
r = 1;
else {
if (nfld[0] <= kpos[0])
error("missing key field in file1", (char *) 0);
if (nfld[1] <= kpos[1])
error("missing key field in file2", (char *) 0);
r = strcmp(fld[0][kpos[0]], fld[1][kpos[1]]);
}
return r;
}
void match()
{
long p;
if (!vflag) (*outfun) (F1 | F2);
p = ftell(fp[1]);
nfld[1] = get2(0); /* check key order */
if (nfld[1] == 0 || strcmp(fld[0][kpos[0]], fld[1][kpos[1]])) {
nfld[0] = get1();
if (strcmp(fld[0][kpos[0]], oldkey) == 0) {
fseek(fp[1], head, 0); /* re-do from head */
nfld[1] = get2(1); /* don't check key order */
} else
head = p; /* mark here */
}
}
void f1_only()
{
if ((aflag & F1) || (vflag & F1)) (*outfun) (F1);
nfld[0] = get1();
}
void f2_only()
{
if ((aflag & F2) || (vflag & F2)) (*outfun) (F2);
head = ftell(fp[1]); /* mark */
nfld[1] = get2(0); /* check key order */
}
void output(f)
{ /* default output form */
if (f & F1)
fputs(fld[0][kpos[0]], stdout);
else
fputs(fld[1][kpos[1]], stdout);
if (f & F1) outfld(0);
if (f & F2) outfld(1);
fputc('\n', stdout);
}
void outfld(file)
{ /* output all fields except key_field */
register int i;
int k, n;
k = kpos[file];
n = nfld[file];
for (i = 0; i < n; i++)
if (i != k) {
fputc(SEP, stdout);
fputs(fld[file][i], stdout);
}
}
void outputf(f)
{ /* output by '-o list' */
int i, j, k;
register char *s;
for (i = k = 0; i < nout; i++) {
j = olist[i].o_file;
if ((f & (j + 1)) && (olist[i].o_field < nfld[j]))
s = fld[j][olist[i].o_field];
else
s = es;
if (s) {
if (k++) fputc(SEP, stdout);
fputs(s, stdout);
}
}
fputc('\n', stdout);
}
int get1()
{ /* read file1 */
int r;
static char oldkey1[BUFSIZ];
if (fld[0][kpos[0]]) {
strcpy(oldkey, fld[0][kpos[0]]); /* save previous key for control */
}
r = getrec(0);
if (r) {
if (strcmp(oldkey1, fld[0][kpos[0]]) > 0)
error("file1 is not sorted", (char *) 0);
strcpy(oldkey1, fld[0][kpos[0]]); /* save prev key for sort check */
}
return r;
}
int get2(back)
{ /* read file2 */
static char oldkey2[BUFSIZ];
int r;
r = getrec(1);
if (r) {
if (!back && strcmp(oldkey2, fld[1][kpos[1]]) > 0)
error("file2 is not sorted", (char *) 0);
strcpy(oldkey2, fld[1][kpos[1]]); /* save prev key for sort check */
}
return r;
}
int getrec(file)
{ /* read one line to split it */
if (fgets(buf[file], BUFSIZ, fp[file]) == (char *) 0)
*buf[file] = '\0';
else if (*buf[file] == '\n' || *buf[file] == '\r')
error("null line in file%s", file ? "1" : "0");
return split(file);
}
int split(file)
{ /* setup fields */
register int n;
register char *s, *t;
for (n = 0, s = buf[file]; *s && *s != '\n' && *s != '\r';) {
if (sep) {
for (t = s; *s && *s != sep && *s != '\n' && *s != '\r'; s++);
} else {
while (*s == ' ' || *s == '\t')
s++; /* skip leading white space */
for (t = s; *s && *s != ' ' && *s != '\t'
&& *s != '\n' && *s != '\r'; s++);
/* We will treat trailing white space as NULL field */
}
if (*s) *s++ = '\0';
fld[file][n++] = t;
if (n == MAXFLD) error("too many filed in file%s", file ? "1" : "0");
}
fld[file][n] = (char *) 0;
return n;
}
FILE *efopen(file, mode)
char *file, *mode;
{
FILE *fp;
if ((fp = fopen(file, mode)) == (FILE *) 0) error("can't open %s", file);
return fp;
}
void error(s, t)
char *s, *t;
{
fprintf(stderr, "%s: ", cmd);
fprintf(stderr, s, t);
fprintf(stderr, "\n");
exit(1);
}

View file

@ -7,7 +7,7 @@ MAN= ash.1 at.1 banner.1 basename.1 \
env.1 expand.1 expr.1 factor.1 \
finger.1 flexdoc.1 fmt.1 fold.1 format.1 fortune.1 \
fsck.mfs.1 head.1 host.1 hostaddr.1 ifdef.1 \
install.1 isodir.1 isoinfo.1 isoread.1 join.1 kill.1 \
install.1 isodir.1 isoinfo.1 isoread.1 kill.1 \
last.1 loadfont.1 loadkeys.1 logger.1 \
look.1 lp.1 ls.1 lspci.1 mail.1 \
mesg.1 mixer.1 mkfs.mfs.1 \

View file

@ -1,119 +0,0 @@
.\" @(#)join.1 6.1 (Berkeley) 4/29/85
.\"
.TH JOIN 1 "April 29, 1985"
.AT 3
.SH NAME
join \- relational database operator
.SH SYNOPSIS
.B join
.RB [ \-a\fIn ]
.RB [ \-e
.IR s ]
.RB [ \-o
.IR list ]
.RB [ \-t\fIc ]
file1 file2
.SH DESCRIPTION
.B Join
forms, on the standard output,
a join
of the two relations specified by the lines of
.I file1
and
.IR file2 .
If
.I file1
is `\-', the standard input is used.
.PP
.I File1
and
.I file2
must be sorted in increasing ASCII collating
sequence on the fields
on which they are to be joined,
normally the first in each line.
.PP
There is one line in the output
for each pair of lines in
.I file1
and
.I file2
that have identical join fields.
The output line normally consists of the common field,
then the rest of the line from
.IR file1 ,
then the rest of the line from
.IR file2 .
.PP
Fields are normally separated by blank, tab or newline.
In this case, multiple separators count as one, and
leading separators are discarded.
.PP
These options are recognized:
.TP
.BI \-a n
In addition to the normal output,
produce a line for each unpairable line in file
.IR n ,
where
.I n
is 1 or 2.
.TP
.BI \-e " s"
Replace empty output fields by string
.IR s .
.ig
.TP
.BI \-j "n m"
Join on the
.IR m th
field of file
.IR n .
If
.I n
is missing, use the
.IR m th
field in each file.
..
.TP
.BI \-o " list"
Each output line comprises the fields specified in
.IR list ,
each element of which has the form
.IR n . m ,
where
.I n
is a file number and
.I m
is a field number.
.PP
.TP
.BI \-t c
Use character
.I c
as a separator (tab character).
Every appearance of
.I c
in a line is significant.
.SH "SEE ALSO"
.BR sort (1),
.BR comm (1),
.BR awk (1x).
.SH BUGS
With default field separation,
the collating sequence is that of
.BR "sort \-b" ;
with
.BR \-t ,
the sequence is that of a plain sort.
.PP
The conventions of
.BR join ,
.BR sort ,
.BR comm ,
.BR uniq ,
.BR look
and
.BR awk (1x)
are wildly incongruous.
.\" ref. to awk(9) man page corrected -- ASW 2005-01-15

View file

@ -53,6 +53,7 @@
2012/10/17 12:00:00,tools/gmp
2012/10/17 12:00:00,tools/headerlist
2012/10/17 12:00:00,tools/host-mkdep
2012/10/17 12:00:00,tools/join
2012/10/17 12:00:00,tools/lex
2012/10/17 12:00:00,tools/lorder
2012/10/17 12:00:00,tools/Makefile
@ -68,6 +69,7 @@
2012/10/17 12:00:00,tools/tic
2012/10/17 12:00:00,usr.bin/gzip/Makefile
2012/10/17 12:00:00,usr.bin/indent
2012/10/17 12:00:00,usr.bin/join
2012/10/17 12:00:00,usr.bin/lorder
2012/10/17 12:00:00,usr.bin/Makefile
2012/10/17 12:00:00,usr.bin/Makefile.inc

View file

@ -60,7 +60,7 @@ LINT_BITS= lint lint2
# Dependencies in SUBDIR below ordered to maximize parallel ability.
SUBDIR= host-mkdep .WAIT compat .WAIT \
binstall .WAIT mktemp .WAIT sed .WAIT \
genassym \
genassym join \
lorder makewhatis mkdep mtree nbperf .WAIT \
m4 \
.WAIT mkfs.mfs \

6
tools/join/Makefile Normal file
View file

@ -0,0 +1,6 @@
# $NetBSD: Makefile,v 1.1 2007/01/14 16:17:29 apb Exp $
HOSTPROGNAME= ${_TOOL_PREFIX}join
HOST_SRCDIR= usr.bin/join
.include "${.CURDIR}/../Makefile.host"

View file

@ -10,7 +10,7 @@ SUBDIR= \
\
du \
genassym \
indent \
indent join \
ldd \
login lorder m4 \
make man \

6
usr.bin/join/Makefile Normal file
View file

@ -0,0 +1,6 @@
# $NetBSD: Makefile,v 1.4 1997/10/19 03:32:11 lukem Exp $
# from: @(#)Makefile 8.1 (Berkeley) 6/6/93
PROG= join
.include <bsd.prog.mk>

209
usr.bin/join/join.1 Normal file
View file

@ -0,0 +1,209 @@
.\" $NetBSD: join.1,v 1.13 2012/04/08 22:00:39 wiz Exp $
.\"
.\" Copyright (c) 1990, 1993
.\" The Regents of the University of California. All rights reserved.
.\"
.\" This code is derived from software contributed to Berkeley by
.\" the Institute of Electrical and Electronics Engineers, Inc.
.\"
.\" Redistribution and use in source and binary forms, with or without
.\" modification, are permitted provided that the following conditions
.\" are met:
.\" 1. Redistributions of source code must retain the above copyright
.\" notice, this list of conditions and the following disclaimer.
.\" 2. Redistributions in binary form must reproduce the above copyright
.\" notice, this list of conditions and the following disclaimer in the
.\" documentation and/or other materials provided with the distribution.
.\" 3. Neither the name of the University nor the names of its contributors
.\" may be used to endorse or promote products derived from this software
.\" without specific prior written permission.
.\"
.\" THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
.\" ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
.\" IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
.\" ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
.\" FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
.\" DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
.\" OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
.\" HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
.\" LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
.\" OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
.\" SUCH DAMAGE.
.\"
.\" from: @(#)join.1 8.3 (Berkeley) 4/28/95
.\" $NetBSD: join.1,v 1.13 2012/04/08 22:00:39 wiz Exp $
.\"
.Dd April 28, 1995
.Dt JOIN 1
.Os
.Sh NAME
.Nm join
.Nd relational database operator
.Sh SYNOPSIS
.Nm
.Op Fl a Ar file_number | Fl v Ar file_number
.Op Fl e Ar string
.Op Fl j Ar file_number field
.Op Fl o Ar list
.Op Fl t Ar char
.Op Fl \&1 Ar field
.Op Fl \&2 Ar field
.Ar file1 file2
.Sh DESCRIPTION
The join utility performs an ``equality join'' on the specified files
and writes the result to the standard output.
The ``join field'' is the field in each file by which the files are compared.
The first field in each line is used by default.
There is one line in the output for each pair of lines in
.Ar file1
and
.Ar file2
which have identical join fields.
Each output line consists of the join field, the remaining fields from
.Ar file1
and then the remaining fields from
.Ar file2 .
.Pp
The default field separators are tab and space characters.
In this case, multiple tabs and spaces count as a single field separator,
and leading tabs and spaces are ignored.
The default output field separator is a single space character.
.Pp
Many of the options use file and field numbers.
Both file numbers and field numbers are 1 based, i.e. the first file on
the command line is file number 1 and the first field is field number 1.
The following options are available:
.Bl -tag -width Fl
.It Fl a Ar file_number
In addition to the default output, produce a line for each unpairable
line in file
.Ar file_number .
(The argument to
.Fl a
must not be preceded by a space; see the
.Sx COMPATIBILITY
section.)
.It Fl e Ar string
Replace empty output fields with
.Ar string .
.It Fl o Ar list
The
.Fl o
option specifies the fields that will be output from each file for
each line with matching join fields.
Each element of
.Ar list
has the form
.Ql file_number.field ,
where
.Ar file_number
is a file number and
.Ar field
is a field number.
The elements of list must be either comma (``,'') or whitespace separated.
(The latter requires quoting to protect it from the shell, or, a simpler
approach is to use multiple
.Fl o
options.)
.It Fl t Ar char
Use character
.Ar char
as a field delimiter for both input and output.
Every occurrence of
.Ar char
in a line is significant.
.It Fl v Ar file_number
Do not display the default output, but display a line for each unpairable
line in file
.Ar file_number .
The options
.Fl v Ar 1
and
.Fl v Ar 2
may be specified at the same time.
.It Fl 1 Ar field
Join on the
.Ar field Ns 'th
field of file 1.
.It Fl 2 Ar field
Join on the
.Ar field Ns 'th
field of file 2.
.El
.Pp
When the default field delimiter characters are used, the files to be joined
should be ordered in the collating sequence of
.Xr sort 1 ,
using the
.Fl b
option, on the fields on which they are to be joined, otherwise
.Nm
may not report all field matches.
When the field delimiter characters are specified by the
.Fl t
option, the collating sequence should be the same as
.Xr sort 1
without the
.Fl b
option.
.Pp
If one of the arguments
.Ar file1
or
.Ar file2
is ``-'', the standard input is used.
.Pp
The
.Nm
utility exits 0 on success, and \*[Gt]0 if an error occurs.
.Sh COMPATIBILITY
For compatibility with historic versions of
.Nm ,
the following options are available:
.Bl -tag -width Fl
.It Fl a
In addition to the default output, produce a line for each unpairable line
in both file 1 and file 2.
(To distinguish between this and
.Fl a Ar file_number ,
.Nm
currently requires that the latter not include any white space.)
.It Fl j1 Ar field
Join on the
.Ar field Ns 'th
field of file 1.
.It Fl j2 Ar field
Join on the
.Ar field Ns 'th
field of file 2.
.It Fl j Ar field
Join on the
.Ar field Ns 'th
field of both file 1 and file 2.
.It Fl o Ar list ...
Historical implementations of
.Nm
permitted multiple arguments to the
.Fl o
option.
These arguments were of the form ``file_number.field_number'' as described
for the current
.Fl o
option.
This has obvious difficulties in the presence of files named ``1.2''.
.El
.Pp
These options are available only so historic shell scripts don't require
modification and should not be used.
.Sh SEE ALSO
.Xr awk 1 ,
.Xr comm 1 ,
.Xr paste 1 ,
.Xr sort 1 ,
.Xr uniq 1
.Sh STANDARDS
The
.Nm
command is expected to be
.St -p1003.2
compatible.

637
usr.bin/join/join.c Normal file
View file

@ -0,0 +1,637 @@
/* $NetBSD: join.c,v 1.31 2011/09/04 20:27:52 joerg Exp $ */
/*-
* Copyright (c) 1991 The Regents of the University of California.
* All rights reserved.
*
* This code is derived from software contributed to Berkeley by
* Steve Hayman of Indiana University, Michiro Hikida and David
* Goodenough.
*
* Redistribution and use in source and binary forms, with or without
* modification, are permitted provided that the following conditions
* are met:
* 1. Redistributions of source code must retain the above copyright
* notice, this list of conditions and the following disclaimer.
* 2. Redistributions in binary form must reproduce the above copyright
* notice, this list of conditions and the following disclaimer in the
* documentation and/or other materials provided with the distribution.
* 3. Neither the name of the University nor the names of its contributors
* may be used to endorse or promote products derived from this software
* without specific prior written permission.
*
* THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
* ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
* IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
* ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
* FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
* DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
* OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
* HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
* LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
* OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
* SUCH DAMAGE.
*/
#if HAVE_NBTOOL_CONFIG_H
#include "nbtool_config.h"
#endif
#include <sys/cdefs.h>
#ifndef lint
__COPYRIGHT("@(#) Copyright (c) 1991\
The Regents of the University of California. All rights reserved.");
#endif /* not lint */
#ifndef lint
#if 0
static char sccsid[] = "from: @(#)join.c 5.1 (Berkeley) 11/18/91";
#else
__RCSID("$NetBSD: join.c,v 1.31 2011/09/04 20:27:52 joerg Exp $");
#endif
#endif /* not lint */
#include <sys/types.h>
#include <ctype.h>
#include <err.h>
#include <errno.h>
#include <stdio.h>
#include <stdlib.h>
#include <string.h>
#include <unistd.h>
/*
* There's a structure per input file which encapsulates the state of the
* file. We repeatedly read lines from each file until we've read in all
* the consecutive lines from the file with a common join field. Then we
* compare the set of lines with an equivalent set from the other file.
*/
typedef struct {
char *line; /* line */
u_long linealloc; /* line allocated count */
char **fields; /* line field(s) */
u_long fieldcnt; /* line field(s) count */
u_long fieldalloc; /* line field(s) allocated count */
} LINE;
static char nolineline[1] = { '\0' };
static LINE noline = {nolineline, 0, 0, 0, 0}; /* arg for outfield if no line to output */
typedef struct {
FILE *fp; /* file descriptor */
u_long joinf; /* join field (-1, -2, -j) */
int unpair; /* output unpairable lines (-a) */
int number; /* 1 for file 1, 2 for file 2 */
LINE *set; /* set of lines with same field */
u_long pushback; /* line on the stack */
u_long setcnt; /* set count */
u_long setalloc; /* set allocated count */
} INPUT;
static INPUT input1 = { NULL, 0, 0, 1, NULL, -1, 0, 0, },
input2 = { NULL, 0, 0, 2, NULL, -1, 0, 0, };
typedef struct {
u_long fileno; /* file number */
u_long fieldno; /* field number */
} OLIST;
static OLIST *olist; /* output field list */
static u_long olistcnt; /* output field list count */
static u_long olistalloc; /* output field allocated count */
static int joinout = 1; /* show lines with matched join fields (-v) */
static int needsep; /* need separator character */
static int spans = 1; /* span multiple delimiters (-t) */
static char *empty; /* empty field replacement string (-e) */
static const char *tabchar = " \t"; /* delimiter characters (-t) */
static int cmp(LINE *, u_long, LINE *, u_long);
__dead static void enomem(void);
static void fieldarg(char *);
static void joinlines(INPUT *, INPUT *);
static void obsolete(char **);
static void outfield(LINE *, u_long);
static void outoneline(INPUT *, LINE *);
static void outtwoline(INPUT *, LINE *, INPUT *, LINE *);
static void slurp(INPUT *);
__dead static void usage(void);
int
main(int argc, char *argv[])
{
INPUT *F1, *F2;
int aflag, ch, cval, vflag;
char *end;
F1 = &input1;
F2 = &input2;
aflag = vflag = 0;
obsolete(argv);
while ((ch = getopt(argc, argv, "\01a:e:j:1:2:o:t:v:")) != -1) {
switch (ch) {
case '\01':
aflag = 1;
F1->unpair = F2->unpair = 1;
break;
case '1':
if ((F1->joinf = strtol(optarg, &end, 10)) < 1) {
warnx("-1 option field number less than 1");
usage();
}
if (*end) {
warnx("illegal field number -- %s", optarg);
usage();
}
--F1->joinf;
break;
case '2':
if ((F2->joinf = strtol(optarg, &end, 10)) < 1) {
warnx("-2 option field number less than 1");
usage();
}
if (*end) {
warnx("illegal field number -- %s", optarg);
usage();
}
--F2->joinf;
break;
case 'a':
aflag = 1;
switch(strtol(optarg, &end, 10)) {
case 1:
F1->unpair = 1;
break;
case 2:
F2->unpair = 1;
break;
default:
warnx("-a option file number not 1 or 2");
usage();
break;
}
if (*end) {
warnx("illegal file number -- %s", optarg);
usage();
}
break;
case 'e':
empty = optarg;
break;
case 'j':
if ((F1->joinf = F2->joinf =
strtol(optarg, &end, 10)) < 1) {
warnx("-j option field number less than 1");
usage();
}
if (*end) {
warnx("illegal field number -- %s", optarg);
usage();
}
--F1->joinf;
--F2->joinf;
break;
case 'o':
fieldarg(optarg);
break;
case 't':
spans = 0;
if (strlen(tabchar = optarg) != 1) {
warnx("illegal tab character specification");
usage();
}
break;
case 'v':
vflag = 1;
joinout = 0;
switch(strtol(optarg, &end, 10)) {
case 1:
F1->unpair = 1;
break;
case 2:
F2->unpair = 1;
break;
default:
warnx("-v option file number not 1 or 2");
usage();
break;
}
if (*end) {
warnx("illegal file number -- %s", optarg);
usage();
}
break;
case '?':
default:
usage();
}
}
argc -= optind;
argv += optind;
if (aflag && vflag)
errx(1, "-a and -v options mutually exclusive");
if (argc != 2)
usage();
/* Open the files; "-" means stdin. */
if (!strcmp(*argv, "-"))
F1->fp = stdin;
else if ((F1->fp = fopen(*argv, "r")) == NULL)
err(1, "%s", *argv);
++argv;
if (!strcmp(*argv, "-"))
F2->fp = stdin;
else if ((F2->fp = fopen(*argv, "r")) == NULL)
err(1, "%s", *argv);
if (F1->fp == stdin && F2->fp == stdin)
errx(1, "only one input file may be stdin");
slurp(F1);
slurp(F2);
while (F1->setcnt && F2->setcnt) {
cval = cmp(F1->set, F1->joinf, F2->set, F2->joinf);
if (cval == 0) {
/* Oh joy, oh rapture, oh beauty divine! */
if (joinout)
joinlines(F1, F2);
slurp(F1);
slurp(F2);
} else if (cval < 0) {
/* File 1 takes the lead... */
if (F1->unpair)
joinlines(F1, NULL);
slurp(F1);
} else {
/* File 2 takes the lead... */
if (F2->unpair)
joinlines(F2, NULL);
slurp(F2);
}
}
/*
* Now that one of the files is used up, optionally output any
* remaining lines from the other file.
*/
if (F1->unpair)
while (F1->setcnt) {
joinlines(F1, NULL);
slurp(F1);
}
if (F1->fp != stdin)
fclose(F1->fp);
if (F2->unpair)
while (F2->setcnt) {
joinlines(F2, NULL);
slurp(F2);
}
if (F2->fp != stdin)
fclose(F2->fp);
return 0;
}
static void
slurp(INPUT *F)
{
LINE *lp;
LINE tmp;
LINE *nline;
size_t len;
u_long cnt;
char *bp, *fieldp;
u_long nsize;
/*
* Read all of the lines from an input file that have the same
* join field.
*/
for (F->setcnt = 0;; ++F->setcnt) {
/*
* If we're out of space to hold line structures, allocate
* more. Initialize the structure so that we know that this
* is new space.
*/
if (F->setcnt == F->setalloc) {
cnt = F->setalloc;
if (F->setalloc == 0)
nsize = 64;
else
nsize = F->setalloc << 1;
if ((nline = realloc(F->set,
nsize * sizeof(LINE))) == NULL)
enomem();
F->set = nline;
F->setalloc = nsize;
memset(F->set + cnt, 0,
(F->setalloc - cnt) * sizeof(LINE));
}
/*
* Get any pushed back line, else get the next line. Allocate
* space as necessary. If taking the line from the stack swap
* the two structures so that we don't lose the allocated space.
* This could be avoided by doing another level of indirection,
* but it's probably okay as is.
*/
lp = &F->set[F->setcnt];
if (F->pushback != (u_long)-1) {
tmp = F->set[F->setcnt];
F->set[F->setcnt] = F->set[F->pushback];
F->set[F->pushback] = tmp;
F->pushback = (u_long)-1;
continue;
}
if ((bp = fgetln(F->fp, &len)) == NULL)
return;
if (lp->linealloc <= len + 1) {
char *n;
if (lp->linealloc == 0)
nsize = 128;
else
nsize = lp->linealloc;
while (nsize <= len + 1)
nsize <<= 1;
if ((n = realloc(lp->line,
nsize * sizeof(char))) == NULL)
enomem();
lp->line = n;
lp->linealloc = nsize;
}
memmove(lp->line, bp, len);
/* Replace trailing newline, if it exists. */
if (bp[len - 1] == '\n')
lp->line[len - 1] = '\0';
else
lp->line[len] = '\0';
bp = lp->line;
/* Split the line into fields, allocate space as necessary. */
lp->fieldcnt = 0;
while ((fieldp = strsep(&bp, tabchar)) != NULL) {
if (spans && *fieldp == '\0')
continue;
if (lp->fieldcnt == lp->fieldalloc) {
char **n;
if (lp->fieldalloc == 0)
nsize = 16;
else
nsize = lp->fieldalloc << 1;
if ((n = realloc(lp->fields,
nsize * sizeof(char *))) == NULL)
enomem();
lp->fields = n;
lp->fieldalloc = nsize;
}
lp->fields[lp->fieldcnt++] = fieldp;
}
/* See if the join field value has changed. */
if (F->setcnt && cmp(lp, F->joinf, lp - 1, F->joinf)) {
F->pushback = F->setcnt;
break;
}
}
}
static int
cmp(LINE *lp1, u_long fieldno1, LINE *lp2, u_long fieldno2)
{
if (lp1->fieldcnt <= fieldno1)
return (lp2->fieldcnt <= fieldno2 ? 0 : 1);
if (lp2->fieldcnt <= fieldno2)
return (-1);
return (strcmp(lp1->fields[fieldno1], lp2->fields[fieldno2]));
}
static void
joinlines(INPUT *F1, INPUT *F2)
{
u_long cnt1, cnt2;
/*
* Output the results of a join comparison. The output may be from
* either file 1 or file 2 (in which case the first argument is the
* file from which to output) or from both.
*/
if (F2 == NULL) {
for (cnt1 = 0; cnt1 < F1->setcnt; ++cnt1)
outoneline(F1, &F1->set[cnt1]);
return;
}
for (cnt1 = 0; cnt1 < F1->setcnt; ++cnt1)
for (cnt2 = 0; cnt2 < F2->setcnt; ++cnt2)
outtwoline(F1, &F1->set[cnt1], F2, &F2->set[cnt2]);
}
static void
outoneline(INPUT *F, LINE *lp)
{
u_long cnt;
/*
* Output a single line from one of the files, according to the
* join rules. This happens when we are writing unmatched single
* lines. Output empty fields in the right places.
*/
if (olist)
for (cnt = 0; cnt < olistcnt; ++cnt) {
if (olist[cnt].fileno == (u_long)F->number)
outfield(lp, olist[cnt].fieldno);
else
outfield(&noline, 1);
}
else
for (cnt = 0; cnt < lp->fieldcnt; ++cnt)
outfield(lp, cnt);
(void)printf("\n");
if (ferror(stdout))
err(1, "stdout");
needsep = 0;
}
static void
outtwoline(INPUT *F1, LINE *lp1, INPUT *F2, LINE *lp2)
{
u_long cnt;
/* Output a pair of lines according to the join list (if any). */
if (olist) {
for (cnt = 0; cnt < olistcnt; ++cnt)
if (olist[cnt].fileno == 1)
outfield(lp1, olist[cnt].fieldno);
else /* if (olist[cnt].fileno == 2) */
outfield(lp2, olist[cnt].fieldno);
} else {
/*
* Output the join field, then the remaining fields from F1
* and F2.
*/
outfield(lp1, F1->joinf);
for (cnt = 0; cnt < lp1->fieldcnt; ++cnt)
if (F1->joinf != cnt)
outfield(lp1, cnt);
for (cnt = 0; cnt < lp2->fieldcnt; ++cnt)
if (F2->joinf != cnt)
outfield(lp2, cnt);
}
(void)printf("\n");
if (ferror(stdout))
err(1, "stdout");
needsep = 0;
}
static void
outfield(LINE *lp, u_long fieldno)
{
if (needsep++)
(void)printf("%c", *tabchar);
if (!ferror(stdout)) {
if (lp->fieldcnt <= fieldno) {
if (empty != NULL)
(void)printf("%s", empty);
} else {
if (*lp->fields[fieldno] == '\0')
return;
(void)printf("%s", lp->fields[fieldno]);
}
}
if (ferror(stdout))
err(1, "stdout");
}
/*
* Convert an output list argument "2.1, 1.3, 2.4" into an array of output
* fields.
*/
static void
fieldarg(char *option)
{
u_long fieldno;
char *end, *token;
OLIST *n;
while ((token = strsep(&option, ", \t")) != NULL) {
if (*token == '\0')
continue;
if ((token[0] != '1' && token[0] != '2') || token[1] != '.')
errx(1, "malformed -o option field");
fieldno = strtol(token + 2, &end, 10);
if (*end)
errx(1, "malformed -o option field");
if (fieldno == 0)
errx(1, "field numbers are 1 based");
if (olistcnt == olistalloc) {
if ((n = realloc(olist,
(olistalloc + 50) * sizeof(OLIST))) == NULL)
enomem();
olist = n;
olistalloc += 50;
}
olist[olistcnt].fileno = token[0] - '0';
olist[olistcnt].fieldno = fieldno - 1;
++olistcnt;
}
}
static void
obsolete(char **argv)
{
size_t len;
char **p, *ap, *t;
while ((ap = *++argv) != NULL) {
/* Return if "--". */
if (ap[0] == '-' && ap[1] == '-')
return;
switch (ap[1]) {
case 'a':
/*
* The original join allowed "-a", which meant the
* same as -a1 plus -a2. POSIX 1003.2, Draft 11.2
* only specifies this as "-a 1" and "a -2", so we
* have to use another option flag, one that is
* unlikely to ever be used or accidentally entered
* on the command line. (Well, we could reallocate
* the argv array, but that hardly seems worthwhile.)
*/
if (ap[2] == '\0')
ap[1] = '\01';
break;
case 'j':
/*
* The original join allowed "-j[12] arg" and "-j arg".
* Convert the former to "-[12] arg". Don't convert
* the latter since getopt(3) can handle it.
*/
switch(ap[2]) {
case '1':
if (ap[3] != '\0')
goto jbad;
ap[1] = '1';
ap[2] = '\0';
break;
case '2':
if (ap[3] != '\0')
goto jbad;
ap[1] = '2';
ap[2] = '\0';
break;
case '\0':
break;
default:
jbad: errx(1, "illegal option -- %s", ap);
usage();
}
break;
case 'o':
/*
* The original join allowed "-o arg arg". Convert to
* "-o arg -o arg".
*/
if (ap[2] != '\0')
break;
for (p = argv + 2; *p; ++p) {
if ((p[0][0] != '1' && p[0][0] != '2') ||
p[0][1] != '.')
break;
len = strlen(*p);
if (len - 2 != strspn(*p + 2, "0123456789"))
break;
if ((t = malloc(len + 3)) == NULL)
enomem();
t[0] = '-';
t[1] = 'o';
memmove(t + 2, *p, len + 1);
*p = t;
}
argv = p - 1;
break;
}
}
}
static void
enomem(void)
{
errx(1, "no memory");
}
static void
usage(void)
{
(void)fprintf(stderr,
"usage: %s [-a fileno | -v fileno] [-e string] [-j fileno field]\n"
" [-o list] [-t char] [-1 field] [-2 field] file1 file2\n",
getprogname());
exit(1);
}