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"
34 gnumeric_open_reader (struct spreadsheet_read_info *gri, struct dictionary **dict)
36 msg (ME, _("Support for %s files was not compiled into this installation of PSPP"), "Gnumeric");
43 #include "data/gnumeric-reader.h"
48 #include <libxml/xmlreader.h>
51 #include "data/case.h"
52 #include "data/casereader-provider.h"
53 #include "data/dictionary.h"
54 #include "data/identifier.h"
55 #include "data/value.h"
56 #include "data/variable.h"
57 #include "libpspp/i18n.h"
58 #include "libpspp/str.h"
60 #include "gl/xalloc.h"
62 static void gnm_file_casereader_destroy (struct casereader *, void *);
64 static struct ccase *gnm_file_casereader_read (struct casereader *, void *);
66 static const struct casereader_class gnm_file_casereader_class =
68 gnm_file_casereader_read,
69 gnm_file_casereader_destroy,
76 STATE_INIT = 0, /* Initial state */
77 STATE_SHEET_START, /* Found the start of a sheet */
78 STATE_SHEET_NAME, /* Found the sheet name */
80 STATE_SHEET_FOUND, /* Found the sheet that we actually want */
81 STATE_CELLS_START, /* Found the start of the cell array */
82 STATE_CELL /* Found a cell */
86 struct gnumeric_reader
90 enum reader_state state;
97 const xmlChar *target_sheet;
98 int target_sheet_index;
105 struct caseproto *proto;
106 struct dictionary *dict;
107 struct ccase *first_case;
108 bool used_first_case;
111 static void process_node (struct gnumeric_reader *r);
115 gnm_file_casereader_destroy (struct casereader *reader UNUSED, void *r_)
117 struct gnumeric_reader *r = r_;
122 xmlFreeTextReader (r->xtr);
124 if ( ! r->used_first_case )
125 case_unref (r->first_case);
127 caseproto_unref (r->proto);
133 process_node (struct gnumeric_reader *r)
135 xmlChar *name = xmlTextReaderName (r->xtr);
137 name = xmlStrdup (_xml ("--"));
140 r->node_type = xmlTextReaderNodeType (r->xtr);
145 if (0 == xmlStrcasecmp (name, _xml("gnm:Sheet")) &&
146 XML_READER_TYPE_ELEMENT == r->node_type)
148 r->state = STATE_SHEET_START;
151 case STATE_SHEET_START:
152 if (0 == xmlStrcasecmp (name, _xml("gnm:Name")) &&
153 XML_READER_TYPE_ELEMENT == r->node_type)
155 r->state = STATE_SHEET_NAME;
157 else if (0 == xmlStrcasecmp (name, _xml("gnm:Name")) &&
158 XML_READER_TYPE_END_ELEMENT == r->node_type)
160 r->state = STATE_INIT;
163 case STATE_SHEET_NAME:
164 if (0 == xmlStrcasecmp (name, _xml("gnm:Name")) &&
165 XML_READER_TYPE_END_ELEMENT == r->node_type)
167 r->state = STATE_SHEET_START;
169 else if (XML_READER_TYPE_TEXT == r->node_type)
172 if ( r->target_sheet != NULL)
174 xmlChar *value = xmlTextReaderValue (r->xtr);
175 if ( 0 == xmlStrcmp (value, r->target_sheet))
176 r->state = STATE_SHEET_FOUND;
179 else if (r->target_sheet_index == r->sheet_index)
181 r->state = STATE_SHEET_FOUND;
185 case STATE_SHEET_FOUND:
186 if (0 == xmlStrcasecmp (name, _xml("gnm:Cells")) &&
187 XML_READER_TYPE_ELEMENT == r->node_type)
189 if (! xmlTextReaderIsEmptyElement (r->xtr))
190 r->state = STATE_CELLS_START;
192 else if (0 == xmlStrcasecmp (name, _xml("gnm:MaxRow")) &&
193 XML_READER_TYPE_ELEMENT == r->node_type)
195 r->state = STATE_MAXROW;
197 else if (0 == xmlStrcasecmp (name, _xml("gnm:Sheet")) &&
198 XML_READER_TYPE_END_ELEMENT == r->node_type)
200 r->state = STATE_INIT;
204 if (0 == xmlStrcasecmp (name, _xml("gnm:MaxRow")) &&
205 XML_READER_TYPE_END_ELEMENT == r->node_type)
207 r->state = STATE_SHEET_FOUND;
209 case STATE_CELLS_START:
210 if (0 == xmlStrcasecmp (name, _xml ("gnm:Cell")) &&
211 XML_READER_TYPE_ELEMENT == r->node_type)
213 xmlChar *attr = NULL;
214 r->state = STATE_CELL;
216 attr = xmlTextReaderGetAttribute (r->xtr, _xml ("Col"));
217 r->col = _xmlchar_to_int (attr);
220 attr = xmlTextReaderGetAttribute (r->xtr, _xml ("Row"));
221 r->row = _xmlchar_to_int (attr);
224 else if (0 == xmlStrcasecmp (name, _xml("gnm:Cells")) &&
225 XML_READER_TYPE_END_ELEMENT == r->node_type)
226 r->state = STATE_SHEET_NAME;
230 if (0 == xmlStrcasecmp (name, _xml("gnm:Cell")) &&
231 XML_READER_TYPE_END_ELEMENT == r->node_type)
232 r->state = STATE_CELLS_START;
243 Sets the VAR of case C, to the value corresponding to the xml string XV
246 convert_xml_string_to_value (struct ccase *c, const struct variable *var,
249 union value *v = case_data_rw (c, var);
252 value_set_missing (v, var_get_width (var));
253 else if ( var_is_alpha (var))
254 value_copy_str_rpad (v, var_get_width (var), xv, ' ');
257 const char *text = CHAR_CAST (const char *, xv);
261 v->f = c_strtod (text, &endptr);
262 if ( errno != 0 || endptr == text)
271 xmlChar *first_value;
275 gnumeric_open_reader (struct spreadsheet_read_info *gri, struct dictionary **dict)
277 unsigned long int vstart = 0;
279 casenumber n_cases = CASENUMBER_MAX;
281 struct var_spec *var_spec = NULL;
284 struct gnumeric_reader *r = NULL;
286 gzFile gz = gzopen (gri->file_name, "r");
290 msg (ME, _("Error opening `%s' for reading as a Gnumeric file: %s."),
291 gri->file_name, strerror (errno));
296 r = xzalloc (sizeof *r);
298 r->xtr = xmlReaderForIO ((xmlInputReadCallback) gzread,
299 (xmlInputCloseCallback) gzclose, gz,
302 if ( r->xtr == NULL )
305 if ( gri->cell_range )
307 if ( ! convert_cell_ref (gri->cell_range,
308 &r->start_col, &r->start_row,
309 &r->stop_col, &r->stop_row))
311 msg (SE, _("Invalid cell range `%s'"),
324 r->state = STATE_INIT;
325 r->target_sheet = BAD_CAST gri->sheet_name;
326 r->target_sheet_index = gri->sheet_index;
327 r->row = r->col = -1;
330 /* Advance to the start of the cells for the target sheet */
331 while ( (r->state != STATE_CELL || r->row < r->start_row )
332 && 1 == (ret = xmlTextReaderRead (r->xtr)))
336 value = xmlTextReaderValue (r->xtr);
338 if ( r->state == STATE_MAXROW && r->node_type == XML_READER_TYPE_TEXT)
340 n_cases = 1 + _xmlchar_to_int (value) ;
346 /* If a range has been given, then use that to calculate the number
348 if ( gri->cell_range)
350 n_cases = MIN (n_cases, r->stop_row - r->start_row + 1);
353 if ( gri->read_names )
359 /* Read in the first row of cells,
360 including the headers if read_names was set */
362 (( r->state == STATE_CELLS_START && r->row <= r->start_row) || r->state == STATE_CELL )
363 && (ret = xmlTextReaderRead (r->xtr))
369 if ( r->row > r->start_row ) break;
371 if ( r->col < r->start_col ||
372 (r->stop_col != -1 && r->col > r->stop_col))
375 idx = r->col - r->start_col;
377 if ( idx >= n_var_specs )
379 n_var_specs = idx + 1 ;
380 var_spec = xrealloc (var_spec, sizeof (*var_spec) * n_var_specs);
381 var_spec [idx].name = NULL;
382 var_spec [idx].width = -1;
383 var_spec [idx].first_value = NULL;
386 if ( r->node_type == XML_READER_TYPE_TEXT )
388 xmlChar *value = xmlTextReaderValue (r->xtr);
389 const char *text = CHAR_CAST (const char *, value);
391 if ( r->row < r->start_row)
393 if ( gri->read_names )
395 var_spec [idx].name = xstrdup (text);
400 var_spec [idx].first_value = xmlStrdup (value);
402 if (-1 == var_spec [idx].width )
403 var_spec [idx].width = (gri->asw == -1) ?
404 ROUND_UP (strlen(text), SPREADSHEET_DEFAULT_WIDTH) : gri->asw;
409 else if ( r->node_type == XML_READER_TYPE_ELEMENT
410 && r->state == STATE_CELL)
412 if ( r->row == r->start_row )
415 xmlTextReaderGetAttribute (r->xtr, _xml ("ValueType"));
417 if ( NULL == attr || 60 != _xmlchar_to_int (attr))
418 var_spec [idx].width = 0;
426 const xmlChar *enc = xmlTextReaderConstEncoding (r->xtr);
429 /* Create the dictionary and populate it */
430 *dict = r->dict = dict_create (CHAR_CAST (const char *, enc));
433 for (i = 0 ; i < n_var_specs ; ++i )
437 /* Probably no data exists for this variable, so allocate a
439 if ( var_spec[i].width == -1 )
440 var_spec[i].width = SPREADSHEET_DEFAULT_WIDTH;
442 name = dict_make_unique_var_name (r->dict, var_spec[i].name, &vstart);
443 dict_create_var (r->dict, name, var_spec[i].width);
447 /* Create the first case, and cache it */
448 r->used_first_case = false;
450 if ( n_var_specs == 0 )
452 msg (MW, _("Selected sheet or range of spreadsheet `%s' is empty."),
457 r->proto = caseproto_ref (dict_get_proto (r->dict));
458 r->first_case = case_create (r->proto);
459 case_set_missing (r->first_case);
461 for ( i = 0 ; i < n_var_specs ; ++i )
463 const struct variable *var = dict_get_var (r->dict, i);
465 convert_xml_string_to_value (r->first_case, var,
466 var_spec[i].first_value);
469 for ( i = 0 ; i < n_var_specs ; ++i )
471 free (var_spec[i].first_value);
472 free (var_spec[i].name);
477 return casereader_create_sequential
481 &gnm_file_casereader_class, r);
485 for ( i = 0 ; i < n_var_specs ; ++i )
487 free (var_spec[i].first_value);
488 free (var_spec[i].name);
492 dict_destroy (*dict);
494 gnm_file_casereader_destroy (NULL, r);
500 /* Reads and returns one case from READER's file. Returns a null
501 pointer on failure. */
502 static struct ccase *
503 gnm_file_casereader_read (struct casereader *reader UNUSED, void *r_)
508 struct gnumeric_reader *r = r_;
509 int current_row = r->row;
511 if ( !r->used_first_case )
513 r->used_first_case = true;
514 return r->first_case;
517 c = case_create (r->proto);
518 case_set_missing (c);
520 while ((r->state == STATE_CELL || r->state == STATE_CELLS_START )
521 && r->row == current_row && (ret = xmlTextReaderRead (r->xtr)))
525 if ( r->col < r->start_col || (r->stop_col != -1 &&
526 r->col > r->stop_col))
529 if ( r->col - r->start_col >= caseproto_get_n_widths (r->proto))
532 if ( r->stop_row != -1 && r->row > r->stop_row)
535 if ( r->node_type == XML_READER_TYPE_TEXT )
537 xmlChar *value = xmlTextReaderValue (r->xtr);
539 const int idx = r->col - r->start_col;
541 const struct variable *var = dict_get_var (r->dict, idx);
543 convert_xml_string_to_value (c, var, value);
560 #endif /* GNM_SUPPORT */