1 /*********************************************************************** 2 * * 3 * This software is part of the ast package * 4 * Copyright (c) 1992-2009 AT&T Intellectual Property * 5 * and is licensed under the * 6 * Common Public License, Version 1.0 * 7 * by AT&T Intellectual Property * 8 * * 9 * A copy of the License is available at * 10 * http://www.opensource.org/licenses/cpl1.0.txt * 11 * (with md5 checksum 059e8cd6165cb4c31e351f2b69388fd9) * 12 * * 13 * Information and Software Systems Research * 14 * AT&T Research * 15 * Florham Park NJ * 16 * * 17 * Glenn Fowler <gsf@research.att.com> * 18 * David Korn <dgk@research.att.com> * 19 * * 20 ***********************************************************************/ 21 #pragma prototyped 22 /* 23 * uniq 24 * 25 * Written by David Korn 26 */ 27 28 static const char usage[] = 29 "[-n?\n@(#)$Id: uniq (AT&T Research) 2009-08-10 $\n]" 30 USAGE_LICENSE 31 "[+NAME?uniq - Report or filter out repeated lines in a file]" 32 "[+DESCRIPTION?\buniq\b reads the input, compares adjacent lines, and " 33 "writes one copy of each input line on the output. The second " 34 "and succeeding copies of the repeated adjacent lines are not " 35 "written.]" 36 "[+?If the output file, \aoutfile\a, is not specified, \buniq\b writes " 37 "to standard output. If no \ainfile\a is given, or if the \ainfile\a " 38 "is \b-\b, \buniq\b reads from standard input with the start of " 39 "the file defined as the current offset.]" 40 "[c:count?Output the number of times each line occurred along with " 41 "the line.]" 42 "[d:repeated|duplicates?Output the first of each duplicate line.]" 43 "[D:all-repeated?Output all duplicate lines as a group with an empty " 44 "line delimiter specified by \adelimit\a:]:?[delimit:=none]" 45 "{" 46 "[n:none?Do not delimit duplicate groups.]" 47 "[p:prepend?Prepend an empty line before each group.]" 48 "[s:separate?Separate each group with an empty line.]" 49 "}" 50 "[f:skip-fields]#[fields?\afields\a is the number of fields to skip over " 51 "before checking for uniqueness. A field is the minimal string matching " 52 "the BRE \b[[:blank:]]]]*[^[:blank:]]]]*\b. -\anumber\a is equivalent to " 53 "\b--skip-fields\b=\anumber\a.]" 54 "[i:ignore-case?Ignore case in comparisons.]" 55 "[s:skip-chars]#[chars?\achars\a is the number of characters to skip over " 56 "before checking for uniqueness. If specified along with \b-f\b, " 57 "the first \achars\a after the first \afields\a are ignored. If " 58 "the \achars\a specifies more characters than are on the line, " 59 "an empty string will be used for comparison. +\anumber\a is " 60 "equivalent to \b--skip-chars\b=\anumber\a.]" 61 "[u:unique?Output unique lines.]" 62 "[w:check-chars]#[chars?\achars\a is the number of characters to compare " 63 "after skipping any specified fields and characters.]" 64 "\n" 65 "\n[infile [outfile]]\n" 66 "\n" 67 "[+EXIT STATUS?]{" 68 "[+0?The input file was successfully processed.]" 69 "[+>0?An error occurred.]" 70 "}" 71 "[+SEE ALSO?\bsort\b(1), \bgrep\b(1)]" 72 ; 73 74 #include <cmd.h> 75 76 #define C_FLAG 1 77 #define D_FLAG 2 78 #define U_FLAG 4 79 80 #define CWIDTH 4 81 #define MAXCNT 9999 82 83 typedef int (*Compare_f)(const char*, const char*, size_t); 84 85 static int uniq(Sfio_t *fdin, Sfio_t *fdout, int fields, int chars, int width, int mode, int* all, Compare_f compare) 86 { 87 register int n, f, outsize=0; 88 register char *cp, *ep, *bufp, *outp; 89 char *orecp, *sbufp=0, *outbuff; 90 int reclen,oreclen= -1,count=0,cwidth=0,sep,next; 91 if(mode&C_FLAG) 92 cwidth = CWIDTH+1; 93 while(1) 94 { 95 if(bufp = sfgetr(fdin,'\n',0)) 96 n = sfvalue(fdin); 97 else if(bufp = sfgetr(fdin,'\n',SF_LASTR)) 98 { 99 n = sfvalue(fdin); 100 bufp = memcpy(fmtbuf(n + 1), bufp, n); 101 bufp[n++] = '\n'; 102 } 103 else 104 n = 0; 105 if(n) 106 { 107 cp = bufp; 108 ep = cp + n; 109 if(f=fields) 110 while(f-->0 && cp<ep) /* skip over fields */ 111 { 112 while(cp<ep && *cp==' ' || *cp=='\t') 113 cp++; 114 while(cp<ep && *cp!=' ' && *cp!='\t') 115 cp++; 116 } 117 if(chars) 118 cp += chars; 119 if((reclen = n - (cp-bufp)) <=0) 120 { 121 reclen = 1; 122 cp = bufp + sfvalue(fdin)-1; 123 } 124 else if(width >= 0 && width < reclen) 125 reclen = width; 126 } 127 else 128 reclen=-2; 129 if(reclen==oreclen && (!reclen || !(*compare)(cp,orecp,reclen))) 130 { 131 count++; 132 if (!all) 133 continue; 134 next = count; 135 } 136 else 137 { 138 next = 0; 139 if(outsize>0) 140 { 141 if(((mode&D_FLAG)&&count==0) || ((mode&U_FLAG)&&count)) 142 { 143 if(outp!=sbufp) 144 sfwrite(fdout,outp,0); 145 } 146 else 147 { 148 if(cwidth) 149 { 150 if(count<9) 151 { 152 f = 0; 153 while(f < CWIDTH-1) 154 outp[f++] = ' '; 155 outp[f++] = '0' + count + 1; 156 outp[f] = ' '; 157 } 158 else if(count<MAXCNT) 159 { 160 count++; 161 f = CWIDTH; 162 outp[f--] = ' '; 163 do 164 { 165 outp[f--] = '0' + (count % 10); 166 } while (count /= 10); 167 while (f >= 0) 168 outp[f--] = ' '; 169 } 170 else 171 { 172 outsize -= (CWIDTH+1); 173 if(outp!=sbufp) 174 { 175 if(!(sbufp=fmtbuf(outsize))) 176 return(1); 177 memcpy(sbufp,outp+CWIDTH+1,outsize); 178 sfwrite(fdout,outp,0); 179 outp = sbufp; 180 } 181 else 182 outp += CWIDTH+1; 183 sfprintf(fdout,"%4d ",count+1); 184 } 185 } 186 if(sfwrite(fdout,outp,outsize) != outsize) 187 return(1); 188 } 189 } 190 } 191 if(n==0) 192 break; 193 if(count = next) 194 { 195 if(sfwrite(fdout,outp,outsize) != outsize) 196 return(1); 197 if(*all >= 0) 198 *all = 1; 199 sep = 0; 200 } 201 else 202 sep = all && *all > 0; 203 /* save current record */ 204 if (!(outbuff = sfreserve(fdout, 0, 0)) || (outsize = sfvalue(fdout)) < 0) 205 return(1); 206 outp = outbuff; 207 if(outsize < n+cwidth+sep) 208 { 209 /* no room in outp, clear lock and use side buffer */ 210 sfwrite(fdout,outp,0); 211 if(!(sbufp = outp=fmtbuf(outsize=n+cwidth+sep))) 212 return(1); 213 } 214 else 215 outsize = n+cwidth+sep; 216 memcpy(outp+cwidth+sep,bufp,n); 217 if(sep) 218 outp[cwidth] = '\n'; 219 oreclen = reclen; 220 orecp = outp+cwidth+sep + (cp-bufp); 221 } 222 return(0); 223 } 224 225 int 226 b_uniq(int argc, char** argv, void* context) 227 { 228 register int n, mode=0; 229 register char *cp; 230 int fields=0, chars=0, width=-1; 231 Sfio_t *fpin, *fpout; 232 int* all = 0; 233 int sep; 234 Compare_f compare = (Compare_f)memcmp; 235 236 cmdinit(argc, argv, context, ERROR_CATALOG, 0); 237 while (n = optget(argv, usage)) switch (n) 238 { 239 case 'c': 240 mode |= C_FLAG; 241 break; 242 case 'd': 243 mode |= D_FLAG; 244 break; 245 case 'D': 246 mode |= D_FLAG; 247 switch ((int)opt_info.num) 248 { 249 case 'p': 250 sep = 1; 251 break; 252 case 's': 253 sep = 0; 254 break; 255 default: 256 sep = -1; 257 break; 258 } 259 all = &sep; 260 break; 261 case 'i': 262 compare = (Compare_f)strncasecmp; 263 break; 264 case 'u': 265 mode |= U_FLAG; 266 break; 267 case 'f': 268 if(*opt_info.option=='-') 269 fields = opt_info.num; 270 else 271 chars = opt_info.num; 272 break; 273 case 's': 274 chars = opt_info.num; 275 break; 276 case 'w': 277 width = opt_info.num; 278 break; 279 case ':': 280 error(2, "%s", opt_info.arg); 281 break; 282 case '?': 283 error(ERROR_usage(2), "%s", opt_info.arg); 284 break; 285 } 286 argv += opt_info.index; 287 if(all && (mode&C_FLAG)) 288 error(2, "-c and -D are mutually exclusive"); 289 if(error_info.errors) 290 error(ERROR_usage(2), "%s", optusage(NiL)); 291 if((cp = *argv) && (argv++,!streq(cp,"-"))) 292 { 293 if(!(fpin = sfopen(NiL,cp,"r"))) 294 error(ERROR_system(1),"%s: cannot open",cp); 295 } 296 else 297 fpin = sfstdin; 298 if(cp = *argv) 299 { 300 argv++; 301 if(!(fpout = sfopen(NiL,cp,"w"))) 302 error(ERROR_system(1),"%s: cannot create",cp); 303 } 304 else 305 fpout = sfstdout; 306 if(*argv) 307 { 308 error(2, "too many arguments"); 309 error(ERROR_usage(2), "%s", optusage(NiL)); 310 } 311 error_info.errors = uniq(fpin,fpout,fields,chars,width,mode,all,compare); 312 if(fpin!=sfstdin) 313 sfclose(fpin); 314 if(fpout!=sfstdout) 315 sfclose(fpout); 316 return(error_info.errors); 317 } 318 319