1 /* PSPP - a program for statistical analysis.
2 Copyright (C) 2010, 2011 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 3 of the License, or
7 (at your option) any later version.
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, see <http://www.gnu.org/licenses/>. */
19 #include "data/csv-file-writer.h"
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/file-name.h"
37 #include "data/format.h"
38 #include "data/make-file.h"
39 #include "data/missing-values.h"
40 #include "data/settings.h"
41 #include "data/value-labels.h"
42 #include "data/variable.h"
43 #include "libpspp/assertion.h"
44 #include "libpspp/i18n.h"
45 #include "libpspp/message.h"
46 #include "libpspp/str.h"
48 #include "gl/ftoastr.h"
49 #include "gl/minmax.h"
50 #include "gl/unlocked-io.h"
51 #include "gl/xalloc.h"
54 #define _(msgid) gettext (msgid)
55 #define N_(msgid) (msgid)
57 /* A variable in a CSV file. */
60 int width; /* Variable width (0 to 32767). */
61 int case_index; /* Index into case. */
62 struct fmt_spec format; /* Print format. */
63 struct missing_values missing; /* User-missing values, if recoding. */
64 struct val_labs *val_labs; /* Value labels, if any and they are in use. */
67 /* Comma-separated value (CSV) file writer. */
70 struct file_handle *fh; /* File handle. */
71 struct fh_lock *lock; /* Mutual exclusion for file. */
72 FILE *file; /* File stream. */
73 struct replace_file *rf; /* Ticket for replacing output file. */
75 struct csv_writer_options opts;
77 char *encoding; /* Encoding used by variables. */
80 struct csv_var *csv_vars; /* Variables. */
81 size_t n_csv_vars; /* Number of variables. */
84 static const struct casewriter_class csv_file_casewriter_class;
86 static void write_var_names (struct csv_writer *, const struct dictionary *);
88 static bool write_error (const struct csv_writer *);
89 static bool close_writer (struct csv_writer *);
91 /* Initializes OPTS with default options for writing a CSV file. */
93 csv_writer_options_init (struct csv_writer_options *opts)
95 opts->recode_user_missing = false;
96 opts->include_var_names = false;
97 opts->use_value_labels = false;
98 opts->use_print_formats = false;
99 opts->decimal = settings_get_decimal_char (FMT_F);
101 opts->qualifier = '"';
104 /* Opens the CSV file designated by file handle FH for writing cases from
105 dictionary DICT according to the given OPTS.
107 No reference to D is retained, so it may be modified or
108 destroyed at will after this function returns. */
110 csv_writer_open (struct file_handle *fh, const struct dictionary *dict,
111 const struct csv_writer_options *opts)
113 struct csv_writer *w;
116 /* Create and initialize writer. */
117 w = xmalloc (sizeof *w);
125 w->encoding = (dict_get_encoding (dict)
126 ? xstrdup (dict_get_encoding (dict))
129 w->n_csv_vars = dict_get_var_cnt (dict);
130 w->csv_vars = xnmalloc (w->n_csv_vars, sizeof *w->csv_vars);
131 for (i = 0; i < w->n_csv_vars; i++)
133 const struct variable *var = dict_get_var (dict, i);
134 struct csv_var *cv = &w->csv_vars[i];
136 cv->width = var_get_width (var);
137 cv->case_index = var_get_case_index (var);
139 cv->format = *var_get_print_format (var);
140 if (opts->recode_user_missing)
141 mv_copy (&cv->missing, var_get_missing_values (var));
143 mv_init (&cv->missing, cv->width);
145 if (opts->use_value_labels)
146 cv->val_labs = val_labs_clone (var_get_value_labels (var));
151 /* Open file handle as an exclusive writer. */
152 /* TRANSLATORS: this fragment will be interpolated into messages in fh_lock()
153 that identify types of files. */
154 w->lock = fh_lock (fh, FH_REF_FILE, N_("CSV file"), FH_ACC_WRITE, true);
158 /* Create the file on disk. */
159 w->rf = replace_file_start (fh_get_file_name (fh), "w", 0666,
163 msg (ME, _("Error opening `%s' for writing as a system file: %s."),
164 fh_get_file_name (fh), strerror (errno));
168 if (opts->include_var_names)
169 write_var_names (w, dict);
174 return casewriter_create (dict_get_proto (dict),
175 &csv_file_casewriter_class, w);
183 csv_field_needs_quoting (struct csv_writer *w, const char *s, size_t len)
187 for (p = s; p < &s[len]; p++)
188 if (*p == w->opts.qualifier || *p == w->opts.delimiter
189 || *p == '\n' || *p == '\r')
196 csv_output_buffer (struct csv_writer *w, const char *s, size_t len)
198 if (csv_field_needs_quoting (w, s, len))
202 putc (w->opts.qualifier, w->file);
203 for (p = s; p < &s[len]; p++)
205 if (*p == w->opts.qualifier)
206 putc (w->opts.qualifier, w->file);
209 putc (w->opts.qualifier, w->file);
212 fwrite (s, 1, len, w->file);
216 csv_output_string (struct csv_writer *w, const char *s)
218 csv_output_buffer (w, s, strlen (s));
222 write_var_names (struct csv_writer *w, const struct dictionary *d)
226 for (i = 0; i < w->n_csv_vars; i++)
229 putc (w->opts.delimiter, w->file);
230 csv_output_string (w, var_get_name (dict_get_var (d, i)));
232 putc ('\n', w->file);
236 csv_output_format (struct csv_writer *w, const struct csv_var *cv,
237 const union value *value)
239 char *s = data_out (value, w->encoding, &cv->format);
240 struct substring ss = ss_cstr (s);
241 if (cv->format.type != FMT_A)
242 ss_trim (&ss, ss_cstr (" "));
244 ss_rtrim (&ss, ss_cstr (" "));
245 csv_output_buffer (w, ss.string, ss.length);
250 extract_date (double number, int *y, int *m, int *d)
254 calendar_offset_to_gregorian (number / 60. / 60. / 24., y, m, d, &yd);
255 return fmod (number, 60. * 60. * 24.);
259 extract_time (double number, double *H, int *M, int *S)
261 *H = floor (number / 60. / 60.);
262 number = fmod (number, 60. * 60.);
264 *M = floor (number / 60.);
265 number = fmod (number, 60.);
271 csv_write_var__ (struct csv_writer *w, const struct csv_var *cv,
272 const union value *value)
276 label = val_labs_find (cv->val_labs, value);
278 csv_output_string (w, label);
279 else if (cv->width == 0 && value->f == SYSMIS)
280 csv_output_buffer (w, " ", 1);
281 else if (w->opts.use_print_formats)
282 csv_output_format (w, cv, value);
285 char s[MAX (DBL_STRLEN_BOUND, 128)];
287 switch (cv->format.type)
311 dtoastr (s, sizeof s, 0, 0, value->f);
312 if (w->opts.decimal != '.')
314 char *cp = strchr (s, '.');
316 *cp = w->opts.decimal;
334 extract_date (value->f, &y, &m, &d);
335 snprintf (s, sizeof s, "%02d/%02d/%04d", m, d, y);
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",
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);
367 csv_output_format (w, cv, value);
370 case FMT_NUMBER_OF_FORMATS:
373 csv_output_string (w, s);
378 csv_write_var (struct csv_writer *w, const struct csv_var *cv,
379 const union value *value)
381 if (mv_is_value_missing (&cv->missing, value, MV_USER))
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);
391 csv_write_var__ (w, cv, value);
395 csv_write_case (struct csv_writer *w, const struct ccase *c)
399 for (i = 0; i < w->n_csv_vars; i++)
401 const struct csv_var *cv = &w->csv_vars[i];
404 putc (w->opts.delimiter, w->file);
405 csv_write_var (w, cv, case_data_idx (c, cv->case_index));
407 putc ('\n', w->file);
410 /* Writes case C to CSV file W. */
412 csv_file_casewriter_write (struct casewriter *writer, void *w_,
415 struct csv_writer *w = w_;
417 if (ferror (w->file))
419 casewriter_force_error (writer);
424 csv_write_case (w, c);
428 /* Destroys CSV file writer W. */
430 csv_file_casewriter_destroy (struct casewriter *writer, void *w_)
432 struct csv_writer *w = w_;
433 if (!close_writer (w))
434 casewriter_force_error (writer);
437 /* Returns true if an I/O error has occurred on WRITER, false otherwise. */
439 write_error (const struct csv_writer *writer)
441 return ferror (writer->file);
444 /* Closes a CSV file after we're done with it.
445 Returns true if successful, false if an I/O error occurred. */
447 close_writer (struct csv_writer *w)
460 if (fclose (w->file) == EOF)
464 msg (ME, _("An I/O error occurred writing CSV file `%s'."),
465 fh_get_file_name (w->fh));
467 if (ok ? !replace_file_commit (w->rf) : !replace_file_abort (w->rf))
476 for (i = 0; i < w->n_csv_vars; i++)
478 struct csv_var *cv = &w->csv_vars[i];
479 mv_destroy (&cv->missing);
480 val_labs_destroy (cv->val_labs);
489 /* CSV file writer casewriter class. */
490 static const struct casewriter_class csv_file_casewriter_class =
492 csv_file_casewriter_write,
493 csv_file_casewriter_destroy,