src/data/csv-file-writer.c

   1 /* PSPP - a program for statistical analysis.
   2    Copyright (C) 2010, 2011, 2012, 2013, 2014 Free Software Foundation, Inc.
   3
   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 3 of the License, or
   7    (at your option) any later version.
   8
   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.
  13
  14    You should have received a copy of the GNU General Public License
  15    along with this program.  If not, see <http://www.gnu.org/licenses/>. */
  16
  17 #include <config.h>
  18
  19 #include "data/csv-file-writer.h"
  20
  21 #include <ctype.h>
  22 #include <errno.h>
  23 #include <math.h>
  24 #include <stdint.h>
  25 #include <stdlib.h>
  26 #include <sys/stat.h>
  27 #include <time.h>
  28
  29 #include "data/calendar.h"
  30 #include "data/case.h"
  31 #include "data/casewriter-provider.h"
  32 #include "data/casewriter.h"
  33 #include "data/data-out.h"
  34 #include "data/dictionary.h"
  35 #include "data/file-handle-def.h"
  36 #include "data/format.h"
  37 #include "data/make-file.h"
  38 #include "data/missing-values.h"
  39 #include "data/settings.h"
  40 #include "data/value-labels.h"
  41 #include "data/variable.h"
  42 #include "libpspp/assertion.h"
  43 #include "libpspp/i18n.h"
  44 #include "libpspp/message.h"
  45 #include "libpspp/str.h"
  46
  47 #include "gl/ftoastr.h"
  48 #include "gl/minmax.h"
  49 #include "gl/unlocked-io.h"
  50 #include "gl/xalloc.h"
  51
  52 #include "gettext.h"
  53 #define _(msgid) gettext (msgid)
  54 #define N_(msgid) (msgid)
  55
  56 /* A variable in a CSV file. */
  57 struct csv_var
  58   {
  59     int width;                     /* Variable width (0 to 32767). */
  60     int case_index;                /* Index into case. */
  61     struct fmt_spec format;        /* Print format. */
  62     struct missing_values missing; /* User-missing values, if recoding. */
  63     struct val_labs *val_labs;  /* Value labels, if any and they are in use. */
  64   };
  65
  66 /* Comma-separated value (CSV) file writer. */
  67 struct csv_writer
  68   {
  69     struct file_handle *fh;     /* File handle. */
  70     struct fh_lock *lock;       /* Mutual exclusion for file. */
  71     FILE *file;                 /* File stream. */
  72     struct replace_file *rf;    /* Ticket for replacing output file. */
  73
  74     struct csv_writer_options opts;
  75
  76     char *encoding;             /* Encoding used by variables. */
  77
  78     /* Variables. */
  79     struct csv_var *csv_vars;   /* Variables. */
  80     size_t n_csv_vars;         /* Number of variables. */
  81   };
  82
  83 static const struct casewriter_class csv_file_casewriter_class;
  84
  85 static void write_var_names (struct csv_writer *, const struct dictionary *);
  86
  87 static bool write_error (const struct csv_writer *);
  88 static bool close_writer (struct csv_writer *);
  89
  90 /* Initializes OPTS with default options for writing a CSV file. */
  91 void
  92 csv_writer_options_init (struct csv_writer_options *opts)
  93 {
  94   opts->recode_user_missing = false;
  95   opts->include_var_names = false;
  96   opts->use_value_labels = false;
  97   opts->use_print_formats = false;
  98   opts->decimal = settings_get_decimal_char (FMT_F);
  99   opts->delimiter = ',';
 100   opts->qualifier = '"';
 101 }
 102
 103 /* Opens the CSV file designated by file handle FH for writing cases from
 104    dictionary DICT according to the given OPTS.
 105
 106    No reference to D is retained, so it may be modified or
 107    destroyed at will after this function returns. */
 108 struct casewriter *
 109 csv_writer_open (struct file_handle *fh, const struct dictionary *dict,
 110                  const struct csv_writer_options *opts)
 111 {
 112   struct csv_writer *w;
 113   int i;
 114
 115   /* Create and initialize writer. */
 116   w = xmalloc (sizeof *w);
 117   w->fh = fh_ref (fh);
 118   w->lock = NULL;
 119   w->file = NULL;
 120   w->rf = NULL;
 121
 122   w->opts = *opts;
 123
 124   w->encoding = xstrdup (dict_get_encoding (dict));
 125
 126   w->n_csv_vars = dict_get_var_cnt (dict);
 127   w->csv_vars = xnmalloc (w->n_csv_vars, sizeof *w->csv_vars);
 128   for (i = 0; i < w->n_csv_vars; i++)
 129     {
 130       const struct variable *var = dict_get_var (dict, i);
 131       struct csv_var *cv = &w->csv_vars[i];
 132
 133       cv->width = var_get_width (var);
 134       cv->case_index = var_get_case_index (var);
 135
 136       cv->format = *var_get_print_format (var);
 137       if (opts->recode_user_missing)
 138         mv_copy (&cv->missing, var_get_missing_values (var));
 139       else
 140         mv_init (&cv->missing, cv->width);
 141
 142       if (opts->use_value_labels)
 143         cv->val_labs = val_labs_clone (var_get_value_labels (var));
 144       else
 145         cv->val_labs = NULL;
 146     }
 147
 148   /* Open file handle as an exclusive writer. */
 149   /* TRANSLATORS: this fragment will be interpolated into messages in fh_lock()
 150      that identify types of files. */
 151   w->lock = fh_lock (fh, FH_REF_FILE, N_("CSV file"), FH_ACC_WRITE, true);
 152   if (w->lock == NULL)
 153     goto error;
 154
 155   /* Create the file on disk. */
 156   w->rf = replace_file_start (fh, "w", 0666,  &w->file);
 157   if (w->rf == NULL)
 158     {
 159       msg (ME, _("Error opening `%s' for writing as a system file: %s."),
 160            fh_get_file_name (fh), strerror (errno));
 161       goto error;
 162     }
 163
 164   if (opts->include_var_names)
 165     write_var_names (w, dict);
 166
 167   if (write_error (w))
 168     goto error;
 169
 170   return casewriter_create (dict_get_proto (dict),
 171                             &csv_file_casewriter_class, w);
 172
 173 error:
 174   close_writer (w);
 175   return NULL;
 176 }
 177
 178 static bool
 179 csv_field_needs_quoting (struct csv_writer *w, const char *s, size_t len)
 180 {
 181   const char *p;
 182
 183   for (p = s; p < &s[len]; p++)
 184     if (*p == w->opts.qualifier || *p == w->opts.delimiter
 185         || *p == '\n' || *p == '\r')
 186       return true;
 187
 188   return false;
 189 }
 190
 191 static void
 192 csv_output_buffer (struct csv_writer *w, const char *s, size_t len)
 193 {
 194   if (csv_field_needs_quoting (w, s, len))
 195     {
 196       const char *p;
 197
 198       putc (w->opts.qualifier, w->file);
 199       for (p = s; p < &s[len]; p++)
 200         {
 201           /* We are writing the output file in text mode, so transform any
 202              explicit CR-LF line breaks into LF only, to allow the C library to
 203              use correct system-specific new-lines. */
 204           if (*p == '\r' && p[1] == '\n')
 205             continue;
 206
 207           if (*p == w->opts.qualifier)
 208             putc (w->opts.qualifier, w->file);
 209           putc (*p, w->file);
 210         }
 211       putc (w->opts.qualifier, w->file);
 212     }
 213   else
 214     fwrite (s, 1, len, w->file);
 215 }
 216
 217 static void
 218 csv_output_string (struct csv_writer *w, const char *s)
 219 {
 220   csv_output_buffer (w, s, strlen (s));
 221 }
 222
 223 static void
 224 write_var_names (struct csv_writer *w, const struct dictionary *d)
 225 {
 226   size_t i;
 227
 228   for (i = 0; i < w->n_csv_vars; i++)
 229     {
 230       if (i > 0)
 231         putc (w->opts.delimiter, w->file);
 232       csv_output_string (w, var_get_name (dict_get_var (d, i)));
 233     }
 234   putc ('\n', w->file);
 235 }
 236
 237 static void
 238 csv_output_format (struct csv_writer *w, const struct csv_var *cv,
 239                    const union value *value)
 240 {
 241   char *s = data_out (value, w->encoding, &cv->format);
 242   struct substring ss = ss_cstr (s);
 243   if (cv->format.type != FMT_A)
 244     ss_trim (&ss, ss_cstr (" "));
 245   else
 246     ss_rtrim (&ss, ss_cstr (" "));
 247   csv_output_buffer (w, ss.string, ss.length);
 248   free (s);
 249 }
 250
 251 static double
 252 extract_date (double number, int *y, int *m, int *d)
 253 {
 254   int yd;
 255
 256   calendar_offset_to_gregorian (number / 60. / 60. / 24., y, m, d, &yd);
 257   return fmod (number, 60. * 60. * 24.);
 258 }
 259
 260 static void
 261 extract_time (double number, double *H, int *M, int *S)
 262 {
 263   *H = floor (number / 60. / 60.);
 264   number = fmod (number, 60. * 60.);
 265
 266   *M = floor (number / 60.);
 267   number = fmod (number, 60.);
 268
 269   *S = floor (number);
 270 }
 271
 272 static void
 273 csv_write_var__ (struct csv_writer *w, const struct csv_var *cv,
 274                  const union value *value)
 275 {
 276   const char *label;
 277
 278   label = val_labs_find (cv->val_labs, value);
 279   if (label != NULL)
 280     csv_output_string (w, label);
 281   else if (cv->width == 0 && value->f == SYSMIS)
 282     csv_output_buffer (w, " ", 1);
 283   else if (w->opts.use_print_formats)
 284     csv_output_format (w, cv, value);
 285   else
 286     {
 287       char s[MAX (DBL_STRLEN_BOUND, 128)];
 288       char *cp;
 289
 290       switch (cv->format.type)
 291         {
 292         case FMT_F:
 293         case FMT_COMMA:
 294         case FMT_DOT:
 295         case FMT_DOLLAR:
 296         case FMT_PCT:
 297         case FMT_E:
 298         case FMT_CCA:
 299         case FMT_CCB:
 300         case FMT_CCC:
 301         case FMT_CCD:
 302         case FMT_CCE:
 303         case FMT_N:
 304         case FMT_Z:
 305         case FMT_P:
 306         case FMT_PK:
 307         case FMT_IB:
 308         case FMT_PIB:
 309         case FMT_PIBHEX:
 310         case FMT_RB:
 311         case FMT_RBHEX:
 312         case FMT_WKDAY:
 313         case FMT_MONTH:
 314           dtoastr (s, sizeof s, 0, 0, value->f);
 315           cp = strpbrk (s, ".,");
 316           if (cp != NULL)
 317             *cp = w->opts.decimal;
 318           break;
 319
 320         case FMT_DATE:
 321         case FMT_ADATE:
 322         case FMT_EDATE:
 323         case FMT_JDATE:
 324         case FMT_SDATE:
 325         case FMT_QYR:
 326         case FMT_MOYR:
 327         case FMT_WKYR:
 328           if (value->f < 0)
 329             strcpy (s, " ");
 330           else
 331             {
 332               int y, m, d;
 333
 334               extract_date (value->f, &y, &m, &d);
 335               snprintf (s, sizeof s, "%02d/%02d/%04d", m, d, y);
 336             }
 337           break;
 338
 339         case FMT_DATETIME:
 340           if (value->f < 0)
 341             strcpy (s, " ");
 342           else
 343             {
 344               int y, m, d, M, S;
 345               double H;
 346
 347               extract_time (extract_date (value->f, &y, &m, &d), &H, &M, &S);
 348               snprintf (s, sizeof s, "%02d/%02d/%04d %02.0f:%02d:%02d",
 349                         m, d, y, H, M, S);
 350             }
 351           break;
 352
 353         case FMT_TIME:
 354         case FMT_DTIME:
 355           {
 356             double H;
 357             int M, S;
 358
 359             extract_time (fabs (value->f), &H, &M, &S);
 360             snprintf (s, sizeof s, "%s%02.0f:%02d:%02d",
 361                       value->f < 0 ? "-" : "", H, M, S);
 362           }
 363           break;
 364
 365         case FMT_A:
 366         case FMT_AHEX:
 367           csv_output_format (w, cv, value);
 368           return;
 369
 370         case FMT_NUMBER_OF_FORMATS:
 371           NOT_REACHED ();
 372         }
 373       csv_output_string (w, s);
 374     }
 375 }
 376
 377 static void
 378 csv_write_var (struct csv_writer *w, const struct csv_var *cv,
 379                const union value *value)
 380 {
 381   if (mv_is_value_missing (&cv->missing, value, MV_USER))
 382     {
 383       union value missing;
 384
 385       value_init (&missing, cv->width);
 386       value_set_missing (&missing, cv->width);
 387       csv_write_var__ (w, cv, &missing);
 388       value_destroy (&missing, cv->width);
 389     }
 390   else
 391     csv_write_var__ (w, cv, value);
 392 }
 393
 394 static void
 395 csv_write_case (struct csv_writer *w, const struct ccase *c)
 396 {
 397   size_t i;
 398
 399   for (i = 0; i < w->n_csv_vars; i++)
 400     {
 401       const struct csv_var *cv = &w->csv_vars[i];
 402
 403       if (i > 0)
 404         putc (w->opts.delimiter, w->file);
 405       csv_write_var (w, cv, case_data_idx (c, cv->case_index));
 406     }
 407   putc ('\n', w->file);
 408 }
 409 \f
 410 /* Writes case C to CSV file W. */
 411 static void
 412 csv_file_casewriter_write (struct casewriter *writer, void *w_,
 413                            struct ccase *c)
 414 {
 415   struct csv_writer *w = w_;
 416
 417   if (ferror (w->file))
 418     {
 419       casewriter_force_error (writer);
 420       case_unref (c);
 421       return;
 422     }
 423
 424   csv_write_case (w, c);
 425   case_unref (c);
 426 }
 427
 428 /* Destroys CSV file writer W. */
 429 static void
 430 csv_file_casewriter_destroy (struct casewriter *writer, void *w_)
 431 {
 432   struct csv_writer *w = w_;
 433   if (!close_writer (w))
 434     casewriter_force_error (writer);
 435 }
 436
 437 /* Returns true if an I/O error has occurred on WRITER, false otherwise. */
 438 bool
 439 write_error (const struct csv_writer *writer)
 440 {
 441   return ferror (writer->file);
 442 }
 443
 444 /* Closes a CSV file after we're done with it.
 445    Returns true if successful, false if an I/O error occurred. */
 446 bool
 447 close_writer (struct csv_writer *w)
 448 {
 449   size_t i;
 450   bool ok;
 451
 452   if (w == NULL)
 453     return true;
 454
 455   ok = true;
 456   if (w->file != NULL)
 457     {
 458       if (write_error (w))
 459         ok = false;
 460       if (fclose (w->file) == EOF)
 461         ok = false;
 462
 463       if (!ok)
 464         msg (ME, _("An I/O error occurred writing CSV file `%s'."),
 465              fh_get_file_name (w->fh));
 466
 467       if (ok ? !replace_file_commit (w->rf) : !replace_file_abort (w->rf))
 468         ok = false;
 469     }
 470
 471   fh_unlock (w->lock);
 472   fh_unref (w->fh);
 473
 474   free (w->encoding);
 475
 476   for (i = 0; i < w->n_csv_vars; i++)
 477     {
 478       struct csv_var *cv = &w->csv_vars[i];
 479       mv_destroy (&cv->missing);
 480       val_labs_destroy (cv->val_labs);
 481     }
 482
 483   free (w->csv_vars);
 484   free (w);
 485
 486   return ok;
 487 }
 488
 489 /* CSV file writer casewriter class. */
 490 static const struct casewriter_class csv_file_casewriter_class =
 491   {
 492     csv_file_casewriter_write,
 493     csv_file_casewriter_destroy,
 494     NULL,
 495   };