1 /* PSPP - a program for statistical analysis.
2 Copyright (C) 2007, 2008, 2009, 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/>. */
21 #include "data/dataset.h"
22 #include "data/dictionary.h"
23 #include "data/format.h"
24 #include "data/gnumeric-reader.h"
25 #include "data/psql-reader.h"
26 #include "data/settings.h"
27 #include "language/command.h"
28 #include "language/data-io/data-parser.h"
29 #include "language/data-io/data-reader.h"
30 #include "language/data-io/file-handle.h"
31 #include "language/data-io/placement-parser.h"
32 #include "language/lexer/format-parser.h"
33 #include "language/lexer/lexer.h"
34 #include "libpspp/i18n.h"
35 #include "libpspp/message.h"
37 #include "gl/xalloc.h"
40 #define _(msgid) gettext (msgid)
41 #define N_(msgid) (msgid)
43 static int parse_get_gnm (struct lexer *lexer, struct dataset *);
44 static int parse_get_txt (struct lexer *lexer, struct dataset *);
45 static int parse_get_psql (struct lexer *lexer, struct dataset *);
48 cmd_get_data (struct lexer *lexer, struct dataset *ds)
50 lex_force_match (lexer, T_SLASH);
52 if (!lex_force_match_id (lexer, "TYPE"))
55 lex_force_match (lexer, T_EQUALS);
57 if (lex_match_id (lexer, "GNM"))
58 return parse_get_gnm (lexer, ds);
59 else if (lex_match_id (lexer, "TXT"))
60 return parse_get_txt (lexer, ds);
61 else if (lex_match_id (lexer, "PSQL"))
62 return parse_get_psql (lexer, ds);
64 msg (SE, _("Unsupported TYPE %s."), lex_tokcstr (lexer));
69 parse_get_psql (struct lexer *lexer, struct dataset *ds)
71 struct psql_read_info psql;
72 psql.allow_clear = false;
76 ds_init_empty (&psql.sql);
78 lex_force_match (lexer, T_SLASH);
80 if (!lex_force_match_id (lexer, "CONNECT"))
83 lex_force_match (lexer, T_EQUALS);
85 if (!lex_force_string (lexer))
88 psql.conninfo = ss_xstrdup (lex_tokss (lexer));
92 while (lex_match (lexer, T_SLASH) )
94 if ( lex_match_id (lexer, "ASSUMEDSTRWIDTH"))
96 lex_match (lexer, T_EQUALS);
97 psql.str_width = lex_integer (lexer);
100 else if ( lex_match_id (lexer, "BSIZE"))
102 lex_match (lexer, T_EQUALS);
103 psql.bsize = lex_integer (lexer);
106 else if ( lex_match_id (lexer, "UNENCRYPTED"))
108 psql.allow_clear = true;
110 else if (lex_match_id (lexer, "SQL"))
112 lex_match (lexer, T_EQUALS);
113 if ( ! lex_force_string (lexer) )
116 ds_put_substring (&psql.sql, lex_tokss (lexer));
121 struct dictionary *dict = NULL;
122 struct casereader *reader = psql_open_reader (&psql, &dict);
125 proc_set_active_file (ds, reader, dict);
128 ds_destroy (&psql.sql);
129 free (psql.conninfo);
135 ds_destroy (&psql.sql);
136 free (psql.conninfo);
142 parse_get_gnm (struct lexer *lexer, struct dataset *ds)
144 struct gnumeric_read_info gri = {NULL, NULL, NULL, 1, true, -1};
146 lex_force_match (lexer, T_SLASH);
148 if (!lex_force_match_id (lexer, "FILE"))
151 lex_force_match (lexer, T_EQUALS);
153 if (!lex_force_string (lexer))
156 gri.file_name = utf8_to_filename (lex_tokcstr (lexer));
160 while (lex_match (lexer, T_SLASH) )
162 if ( lex_match_id (lexer, "ASSUMEDSTRWIDTH"))
164 lex_match (lexer, T_EQUALS);
165 gri.asw = lex_integer (lexer);
168 else if (lex_match_id (lexer, "SHEET"))
170 lex_match (lexer, T_EQUALS);
171 if (lex_match_id (lexer, "NAME"))
173 if ( ! lex_force_string (lexer) )
176 gri.sheet_name = ss_xstrdup (lex_tokss (lexer));
177 gri.sheet_index = -1;
181 else if (lex_match_id (lexer, "INDEX"))
183 gri.sheet_index = lex_integer (lexer);
189 else if (lex_match_id (lexer, "CELLRANGE"))
191 lex_match (lexer, T_EQUALS);
193 if (lex_match_id (lexer, "FULL"))
195 gri.cell_range = NULL;
197 else if (lex_match_id (lexer, "RANGE"))
199 if ( ! lex_force_string (lexer) )
202 gri.cell_range = ss_xstrdup (lex_tokss (lexer));
208 else if (lex_match_id (lexer, "READNAMES"))
210 lex_match (lexer, T_EQUALS);
212 if ( lex_match_id (lexer, "ON"))
214 gri.read_names = true;
216 else if (lex_match_id (lexer, "OFF"))
218 gri.read_names = false;
225 lex_error (lexer, NULL);
231 struct dictionary *dict = NULL;
232 struct casereader *reader = gnumeric_open_reader (&gri, &dict);
235 proc_set_active_file (ds, reader, dict);
238 free (gri.file_name);
239 free (gri.sheet_name);
240 free (gri.cell_range);
245 free (gri.file_name);
246 free (gri.sheet_name);
247 free (gri.cell_range);
252 set_type (struct data_parser *parser, const char *subcommand,
253 enum data_parser_type type, bool *has_type)
257 data_parser_set_type (parser, type);
260 else if (type != data_parser_get_type (parser))
262 msg (SE, _("%s is allowed only with %s arrangement, but %s arrangement "
263 "was stated or implied earlier in this command."),
265 type == DP_FIXED ? "FIXED" : "DELIMITED",
266 type == DP_FIXED ? "DELIMITED" : "FIXED");
273 parse_get_txt (struct lexer *lexer, struct dataset *ds)
275 struct data_parser *parser = NULL;
276 struct dictionary *dict = dict_create ();
277 struct file_handle *fh = NULL;
278 struct dfm_reader *reader = NULL;
282 enum data_parser_type type;
285 lex_force_match (lexer, T_SLASH);
287 if (!lex_force_match_id (lexer, "FILE"))
289 lex_force_match (lexer, T_EQUALS);
290 fh = fh_parse (lexer, FH_REF_FILE | FH_REF_INLINE);
294 parser = data_parser_create (dict);
296 data_parser_set_type (parser, DP_DELIMITED);
297 data_parser_set_span (parser, false);
298 data_parser_set_quotes (parser, ss_empty ());
299 data_parser_set_empty_line_has_field (parser, true);
303 if (!lex_force_match (lexer, T_SLASH))
306 if (lex_match_id (lexer, "ARRANGEMENT"))
310 lex_match (lexer, T_EQUALS);
311 if (lex_match_id (lexer, "FIXED"))
312 ok = set_type (parser, "ARRANGEMENT=FIXED", DP_FIXED, &has_type);
313 else if (lex_match_id (lexer, "DELIMITED"))
314 ok = set_type (parser, "ARRANGEMENT=DELIMITED",
315 DP_DELIMITED, &has_type);
318 lex_error (lexer, _("expecting %s or %s"), "FIXED", "DELIMITED");
324 else if (lex_match_id (lexer, "FIRSTCASE"))
326 lex_match (lexer, T_EQUALS);
327 if (!lex_force_int (lexer))
329 if (lex_integer (lexer) < 1)
331 msg (SE, _("Value of FIRSTCASE must be 1 or greater."));
334 data_parser_set_skip (parser, lex_integer (lexer) - 1);
337 else if (lex_match_id_n (lexer, "DELCASE", 4))
339 if (!set_type (parser, "DELCASE", DP_DELIMITED, &has_type))
341 lex_match (lexer, T_EQUALS);
342 if (lex_match_id (lexer, "LINE"))
343 data_parser_set_span (parser, false);
344 else if (lex_match_id (lexer, "VARIABLES"))
346 data_parser_set_span (parser, true);
348 /* VARIABLES takes an integer argument, but for no
349 good reason. We just ignore it. */
350 if (!lex_force_int (lexer))
356 lex_error (lexer, _("expecting %s or %s"), "LINE", "VARIABLES");
360 else if (lex_match_id (lexer, "FIXCASE"))
362 if (!set_type (parser, "FIXCASE", DP_FIXED, &has_type))
364 lex_match (lexer, T_EQUALS);
365 if (!lex_force_int (lexer))
367 if (lex_integer (lexer) < 1)
369 msg (SE, _("Value of FIXCASE must be at least 1."));
372 data_parser_set_records (parser, lex_integer (lexer));
375 else if (lex_match_id (lexer, "IMPORTCASES"))
377 lex_match (lexer, T_EQUALS);
378 if (lex_match (lexer, T_ALL))
380 data_parser_set_case_limit (parser, -1);
381 data_parser_set_case_percent (parser, 100);
383 else if (lex_match_id (lexer, "FIRST"))
385 if (!lex_force_int (lexer))
387 if (lex_integer (lexer) < 1)
389 msg (SE, _("Value of FIRST must be at least 1."));
392 data_parser_set_case_limit (parser, lex_integer (lexer));
395 else if (lex_match_id (lexer, "PERCENT"))
397 if (!lex_force_int (lexer))
399 if (lex_integer (lexer) < 1 || lex_integer (lexer) > 100)
401 msg (SE, _("Value of PERCENT must be between 1 and 100."));
404 data_parser_set_case_percent (parser, lex_integer (lexer));
408 else if (lex_match_id_n (lexer, "DELIMITERS", 4))
410 struct string hard_seps = DS_EMPTY_INITIALIZER;
411 const char *soft_seps = "";
415 if (!set_type (parser, "DELIMITERS", DP_DELIMITED, &has_type))
417 lex_match (lexer, T_EQUALS);
419 if (!lex_force_string (lexer))
422 /* XXX should support multibyte UTF-8 characters */
423 s = lex_tokss (lexer);
424 if (ss_match_string (&s, ss_cstr ("\\t")))
425 ds_put_cstr (&hard_seps, "\t");
426 if (ss_match_string (&s, ss_cstr ("\\\\")))
427 ds_put_cstr (&hard_seps, "\\");
428 while ((c = ss_get_byte (&s)) != EOF)
432 ds_put_byte (&hard_seps, c);
433 data_parser_set_soft_delimiters (parser, ss_cstr (soft_seps));
434 data_parser_set_hard_delimiters (parser, ds_ss (&hard_seps));
435 ds_destroy (&hard_seps);
439 else if (lex_match_id (lexer, "QUALIFIERS"))
441 if (!set_type (parser, "QUALIFIERS", DP_DELIMITED, &has_type))
443 lex_match (lexer, T_EQUALS);
445 if (!lex_force_string (lexer))
448 /* XXX should support multibyte UTF-8 characters */
449 if (settings_get_syntax () == COMPATIBLE
450 && ss_length (lex_tokss (lexer)) != 1)
452 msg (SE, _("In compatible syntax mode, the QUALIFIER string "
453 "must contain exactly one character."));
457 data_parser_set_quotes (parser, lex_tokss (lexer));
460 else if (settings_get_syntax () == ENHANCED
461 && lex_match_id (lexer, "ESCAPE"))
462 data_parser_set_quote_escape (parser, true);
463 else if (lex_match_id (lexer, "VARIABLES"))
467 lex_error (lexer, _("expecting %s"), "VARIABLES");
471 lex_match (lexer, T_EQUALS);
474 type = data_parser_get_type (parser);
477 struct fmt_spec input, output;
481 while (type == DP_FIXED && lex_match (lexer, T_SLASH))
483 if (!lex_force_int (lexer))
485 if (lex_integer (lexer) < record)
487 msg (SE, _("The record number specified, %ld, is at or "
488 "before the previous record, %d. Data "
489 "fields must be listed in order of "
490 "increasing record number."),
491 lex_integer (lexer), record);
494 if (lex_integer (lexer) > data_parser_get_records (parser))
496 msg (SE, _("The record number specified, %ld, exceeds "
497 "the number of records per case specified "
499 lex_integer (lexer), data_parser_get_records (parser));
502 record = lex_integer (lexer);
506 if (!lex_force_id (lexer)
507 || !dict_id_is_valid (dict, lex_tokcstr (lexer), true))
509 name = xstrdup (lex_tokcstr (lexer));
512 if (type == DP_DELIMITED)
514 if (!parse_format_specifier (lexer, &input)
515 || !fmt_check_input (&input))
520 if (!parse_column_range (lexer, 0, &fc, &lc, NULL))
522 if (!parse_format_specifier_name (lexer, &input.type))
524 input.w = lc - fc + 1;
526 if (!fmt_check_input (&input))
529 output = fmt_for_output_from_input (&input);
531 v = dict_create_var (dict, name, fmt_var_width (&input));
534 msg (SE, _("%s is a duplicate variable name."), name);
537 var_set_both_formats (v, &output);
539 if (type == DP_DELIMITED)
540 data_parser_add_delimited_field (parser, &input,
541 var_get_case_index (v),
544 data_parser_add_fixed_field (parser, &input, var_get_case_index (v),
549 while (lex_token (lexer) != T_ENDCMD);
551 reader = dfm_open_reader (fh, lexer);
555 data_parser_make_active_file (parser, ds, reader, dict);
560 data_parser_destroy (parser);
564 return CMD_CASCADING_FAILURE;