1 /* PSPP - a program for statistical analysis.
2 Copyright (C) 2007, 2009, 2010, 2011, 2012 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 "libpspp/message.h"
20 #include "libpspp/misc.h"
22 #include "gl/minmax.h"
23 #include "gl/c-strtod.h"
26 #define _(msgid) gettext (msgid)
27 #define N_(msgid) (msgid)
29 #include "spreadsheet-reader.h"
31 #include "c-xvasprintf.h"
36 gnumeric_open_reader (struct spreadsheet_read_info *gri, struct spreadsheet_read_options *opts, struct dictionary **dict)
38 msg (ME, _("Support for %s files was not compiled into this installation of PSPP"), "Gnumeric");
45 #include "data/gnumeric-reader.h"
50 #include <libxml/xmlreader.h>
53 #include "data/case.h"
54 #include "data/casereader-provider.h"
55 #include "data/dictionary.h"
56 #include "data/identifier.h"
57 #include "data/value.h"
58 #include "data/variable.h"
59 #include "libpspp/i18n.h"
60 #include "libpspp/str.h"
62 #include "gl/xalloc.h"
64 static void gnm_file_casereader_destroy (struct casereader *, void *);
66 static struct ccase *gnm_file_casereader_read (struct casereader *, void *);
68 static const struct casereader_class gnm_file_casereader_class =
70 gnm_file_casereader_read,
71 gnm_file_casereader_destroy,
78 STATE_INIT = 0, /* Initial state */
79 STATE_SHEET_START, /* Found the start of a sheet */
80 STATE_SHEET_NAME, /* Found the sheet name */
82 STATE_SHEET_FOUND, /* Found the sheet that we actually want */
83 STATE_CELLS_START, /* Found the start of the cell array */
84 STATE_CELL /* Found a cell */
88 struct gnumeric_reader
92 enum reader_state state;
100 const xmlChar *target_sheet;
101 int target_sheet_index;
108 struct caseproto *proto;
109 struct dictionary *dict;
110 struct ccase *first_case;
111 bool used_first_case;
114 static void process_node (struct gnumeric_reader *r);
118 gnm_file_casereader_destroy (struct casereader *reader UNUSED, void *r_)
120 struct gnumeric_reader *r = r_;
125 xmlFreeTextReader (r->xtr);
127 if ( ! r->used_first_case )
128 case_unref (r->first_case);
130 caseproto_unref (r->proto);
136 process_node (struct gnumeric_reader *r)
138 xmlChar *name = xmlTextReaderName (r->xtr);
140 name = xmlStrdup (_xml ("--"));
143 r->node_type = xmlTextReaderNodeType (r->xtr);
148 if (0 == xmlStrcasecmp (name, _xml("gnm:Sheet")) &&
149 XML_READER_TYPE_ELEMENT == r->node_type)
152 r->state = STATE_SHEET_START;
155 case STATE_SHEET_START:
156 if (0 == xmlStrcasecmp (name, _xml("gnm:Name")) &&
157 XML_READER_TYPE_ELEMENT == r->node_type)
159 r->state = STATE_SHEET_NAME;
162 case STATE_SHEET_NAME:
163 if (0 == xmlStrcasecmp (name, _xml("gnm:Name")) &&
164 XML_READER_TYPE_END_ELEMENT == r->node_type)
166 r->state = STATE_INIT;
168 else if (XML_READER_TYPE_TEXT == r->node_type)
170 if ( r->target_sheet != NULL)
172 xmlChar *value = xmlTextReaderValue (r->xtr);
173 if ( 0 == xmlStrcmp (value, r->target_sheet))
174 r->state = STATE_SHEET_FOUND;
177 else if (r->target_sheet_index == r->sheet_index)
179 r->state = STATE_SHEET_FOUND;
183 case STATE_SHEET_FOUND:
184 if (0 == xmlStrcasecmp (name, _xml("gnm:Cells")) &&
185 XML_READER_TYPE_ELEMENT == r->node_type)
187 r->min_col = INT_MAX;
188 if (! xmlTextReaderIsEmptyElement (r->xtr))
189 r->state = STATE_CELLS_START;
191 else if (0 == xmlStrcasecmp (name, _xml("gnm:MaxRow")) &&
192 XML_READER_TYPE_ELEMENT == r->node_type)
194 r->state = STATE_MAXROW;
196 else if (0 == xmlStrcasecmp (name, _xml("gnm:Sheet")) &&
197 XML_READER_TYPE_END_ELEMENT == r->node_type)
199 r->state = STATE_INIT;
203 if (0 == xmlStrcasecmp (name, _xml("gnm:MaxRow")) &&
204 XML_READER_TYPE_END_ELEMENT == r->node_type)
206 r->state = STATE_SHEET_FOUND;
208 case STATE_CELLS_START:
209 if (0 == xmlStrcasecmp (name, _xml ("gnm:Cell")) &&
210 XML_READER_TYPE_ELEMENT == r->node_type)
212 xmlChar *attr = NULL;
213 r->state = STATE_CELL;
215 attr = xmlTextReaderGetAttribute (r->xtr, _xml ("Col"));
216 r->col = _xmlchar_to_int (attr);
219 if (r->col < r->min_col)
222 attr = xmlTextReaderGetAttribute (r->xtr, _xml ("Row"));
223 r->row = _xmlchar_to_int (attr);
226 else if (0 == xmlStrcasecmp (name, _xml("gnm:Cells")) &&
227 XML_READER_TYPE_END_ELEMENT == r->node_type)
228 r->state = STATE_SHEET_NAME;
232 if (0 == xmlStrcasecmp (name, _xml("gnm:Cell")) &&
233 XML_READER_TYPE_END_ELEMENT == r->node_type)
235 r->state = STATE_CELLS_START;
247 Sets the VAR of case C, to the value corresponding to the xml string XV
250 convert_xml_string_to_value (struct ccase *c, const struct variable *var,
253 union value *v = case_data_rw (c, var);
256 value_set_missing (v, var_get_width (var));
257 else if ( var_is_alpha (var))
258 value_copy_str_rpad (v, var_get_width (var), xv, ' ');
261 const char *text = CHAR_CAST (const char *, xv);
265 v->f = c_strtod (text, &endptr);
266 if ( errno != 0 || endptr == text)
275 xmlChar *first_value;
279 gnumeric_open_reader (const struct spreadsheet_read_info *gri,
280 struct spreadsheet_read_options *opts,
281 struct dictionary **dict)
283 unsigned long int vstart = 0;
285 casenumber n_cases = CASENUMBER_MAX;
287 struct var_spec *var_spec = NULL;
290 struct gnumeric_reader *r = NULL;
292 gzFile gz = gzopen (gri->file_name, "r");
296 msg (ME, _("Error opening `%s' for reading as a Gnumeric file: %s."),
297 gri->file_name, strerror (errno));
302 r = xzalloc (sizeof *r);
304 r->xtr = xmlReaderForIO ((xmlInputReadCallback) gzread,
305 (xmlInputCloseCallback) gzclose, gz,
308 if ( r->xtr == NULL )
311 if ( opts->cell_range )
313 if ( ! convert_cell_ref (opts->cell_range,
314 &r->start_col, &r->start_row,
315 &r->stop_col, &r->stop_row))
317 msg (SE, _("Invalid cell range `%s'"),
330 r->state = STATE_INIT;
331 r->target_sheet = BAD_CAST opts->sheet_name;
332 r->target_sheet_index = opts->sheet_index;
333 r->row = r->col = -1;
336 /* Advance to the start of the cells for the target sheet */
337 while ( (r->state != STATE_CELL || r->row < r->start_row )
338 && 1 == (ret = xmlTextReaderRead (r->xtr)))
342 value = xmlTextReaderValue (r->xtr);
344 if ( r->state == STATE_MAXROW && r->node_type == XML_READER_TYPE_TEXT)
346 n_cases = 1 + _xmlchar_to_int (value) ;
352 /* If a range has been given, then use that to calculate the number
354 if ( opts->cell_range)
356 n_cases = MIN (n_cases, r->stop_row - r->start_row + 1);
359 if ( gri->read_names )
365 /* Read in the first row of cells,
366 including the headers if read_names was set */
368 (( r->state == STATE_CELLS_START && r->row <= r->start_row) || r->state == STATE_CELL )
369 && (ret = xmlTextReaderRead (r->xtr))
375 if ( r->row > r->start_row ) break;
377 if ( r->col < r->start_col ||
378 (r->stop_col != -1 && r->col > r->stop_col))
381 idx = r->col - r->start_col;
383 if ( idx >= n_var_specs )
386 var_spec = xrealloc (var_spec, sizeof (*var_spec) * (idx + 1));
387 for (i = n_var_specs; i <= idx; ++i)
389 var_spec [i].name = NULL;
390 var_spec [i].width = -1;
391 var_spec [i].first_value = NULL;
393 n_var_specs = idx + 1 ;
396 if ( r->node_type == XML_READER_TYPE_TEXT )
398 xmlChar *value = xmlTextReaderValue (r->xtr);
399 const char *text = CHAR_CAST (const char *, value);
401 if ( r->row < r->start_row)
403 if ( gri->read_names )
405 var_spec [idx].name = xstrdup (text);
410 var_spec [idx].first_value = xmlStrdup (value);
412 if (-1 == var_spec [idx].width )
413 var_spec [idx].width = (gri->asw == -1) ?
414 ROUND_UP (strlen(text), SPREADSHEET_DEFAULT_WIDTH) : gri->asw;
419 else if ( r->node_type == XML_READER_TYPE_ELEMENT
420 && r->state == STATE_CELL)
422 if ( r->row == r->start_row )
425 xmlTextReaderGetAttribute (r->xtr, _xml ("ValueType"));
427 if ( NULL == attr || 60 != _xmlchar_to_int (attr))
428 var_spec [idx].width = 0;
436 const xmlChar *enc = xmlTextReaderConstEncoding (r->xtr);
439 /* Create the dictionary and populate it */
440 *dict = r->dict = dict_create (CHAR_CAST (const char *, enc));
443 for (i = 0 ; i < n_var_specs ; ++i )
447 if ( (var_spec[i].name == NULL) && (var_spec[i].first_value == NULL))
450 /* Probably no data exists for this variable, so allocate a
452 if ( var_spec[i].width == -1 )
453 var_spec[i].width = SPREADSHEET_DEFAULT_WIDTH;
455 name = dict_make_unique_var_name (r->dict, var_spec[i].name, &vstart);
456 dict_create_var (r->dict, name, var_spec[i].width);
460 /* Create the first case, and cache it */
461 r->used_first_case = false;
463 if ( n_var_specs == 0 )
465 msg (MW, _("Selected sheet or range of spreadsheet `%s' is empty."),
470 r->proto = caseproto_ref (dict_get_proto (r->dict));
471 r->first_case = case_create (r->proto);
472 case_set_missing (r->first_case);
475 for ( i = 0 ; i < n_var_specs ; ++i )
477 if ( (var_spec[i].name == NULL) && (var_spec[i].first_value == NULL))
480 const struct variable *var = dict_get_var (r->dict, x++);
482 convert_xml_string_to_value (r->first_case, var,
483 var_spec[i].first_value);
486 for ( i = 0 ; i < n_var_specs ; ++i )
488 free (var_spec[i].first_value);
489 free (var_spec[i].name);
495 if (opts->cell_range == NULL)
497 opts->cell_range = c_xasprintf ("%c%d:%c%d",
500 r->stop_col + 'A' + caseproto_get_n_widths (r->proto),
501 r->start_row + n_cases);
504 return casereader_create_sequential
508 &gnm_file_casereader_class, r);
512 for ( i = 0 ; i < n_var_specs ; ++i )
514 free (var_spec[i].first_value);
515 free (var_spec[i].name);
519 dict_destroy (*dict);
522 gnm_file_casereader_destroy (NULL, r);
528 /* Reads and returns one case from READER's file. Returns a null
529 pointer on failure. */
530 static struct ccase *
531 gnm_file_casereader_read (struct casereader *reader UNUSED, void *r_)
536 struct gnumeric_reader *r = r_;
537 int current_row = r->row;
539 if ( !r->used_first_case )
541 r->used_first_case = true;
542 return r->first_case;
545 c = case_create (r->proto);
546 case_set_missing (c);
548 if (r->start_col == -1)
549 r->start_col = r->min_col;
551 while ((r->state == STATE_CELL || r->state == STATE_CELLS_START )
552 && r->row == current_row && (ret = xmlTextReaderRead (r->xtr)))
556 if ( r->col < r->start_col || (r->stop_col != -1 &&
557 r->col > r->stop_col))
560 if ( r->col - r->start_col >= caseproto_get_n_widths (r->proto))
563 if ( r->stop_row != -1 && r->row > r->stop_row)
566 if ( r->node_type == XML_READER_TYPE_TEXT )
568 xmlChar *value = xmlTextReaderValue (r->xtr);
570 const int idx = r->col - r->start_col;
572 const struct variable *var = dict_get_var (r->dict, idx);
574 convert_xml_string_to_value (c, var, value);
591 #endif /* GNM_SUPPORT */