1 /* PSPP - a program for statistical analysis.
2 Copyright (C) 1997-9, 2000, 2006, 2007, 2009, 2010, 2011, 2012, 2013, 2016 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/>. */
25 #include "data/case.h"
26 #include "data/casereader.h"
27 #include "data/data-in.h"
28 #include "data/dataset.h"
29 #include "data/dictionary.h"
30 #include "data/format.h"
31 #include "data/settings.h"
32 #include "data/transformations.h"
33 #include "data/variable.h"
34 #include "language/command.h"
35 #include "language/data-io/data-parser.h"
36 #include "language/data-io/data-reader.h"
37 #include "language/data-io/file-handle.h"
38 #include "language/data-io/inpt-pgm.h"
39 #include "language/data-io/placement-parser.h"
40 #include "language/lexer/format-parser.h"
41 #include "language/lexer/lexer.h"
42 #include "language/lexer/variable-parser.h"
43 #include "libpspp/assertion.h"
44 #include "libpspp/compiler.h"
45 #include "libpspp/i18n.h"
46 #include "libpspp/message.h"
47 #include "libpspp/misc.h"
48 #include "libpspp/pool.h"
49 #include "libpspp/str.h"
52 #include "gl/xalloc.h"
55 #define _(msgid) gettext (msgid)
57 /* DATA LIST transformation data. */
60 struct data_parser *parser; /* Parser. */
61 struct dfm_reader *reader; /* Data file reader. */
62 struct variable *end; /* Variable specified on END subcommand. */
65 static bool parse_fixed (struct lexer *, struct dictionary *,
66 struct pool *, struct data_parser *);
67 static bool parse_free (struct lexer *, struct dictionary *,
68 struct pool *, struct data_parser *);
70 static trns_free_func data_list_trns_free;
71 static trns_proc_func data_list_trns_proc;
74 cmd_data_list (struct lexer *lexer, struct dataset *ds)
76 struct dictionary *dict;
77 struct data_parser *parser;
78 struct dfm_reader *reader;
79 struct variable *end = NULL;
80 struct file_handle *fh = NULL;
81 char *encoding = NULL;
84 enum data_parser_type type;
86 struct pool *tmp_pool;
89 dict = (in_input_program ()
91 : dict_create (get_default_encoding ()));
92 parser = data_parser_create (dict);
95 table = -1; /* Print table if nonzero, -1=undecided. */
98 while (lex_token (lexer) != T_SLASH)
100 if (lex_match_id (lexer, "FILE"))
102 lex_match (lexer, T_EQUALS);
104 fh = fh_parse (lexer, FH_REF_FILE | FH_REF_INLINE, NULL);
108 else if (lex_match_id (lexer, "ENCODING"))
110 lex_match (lexer, T_EQUALS);
111 if (!lex_force_string (lexer))
115 encoding = ss_xstrdup (lex_tokss (lexer));
119 else if (lex_match_id (lexer, "RECORDS"))
121 lex_match (lexer, T_EQUALS);
122 lex_match (lexer, T_LPAREN);
123 if (!lex_force_int (lexer))
125 data_parser_set_records (parser, lex_integer (lexer));
127 lex_match (lexer, T_RPAREN);
129 else if (lex_match_id (lexer, "SKIP"))
131 lex_match (lexer, T_EQUALS);
132 if (!lex_force_int (lexer))
134 data_parser_set_skip (parser, lex_integer (lexer));
137 else if (lex_match_id (lexer, "END"))
139 if (!in_input_program ())
141 msg (SE, _("The %s subcommand may only be used within %s."), "END", "INPUT PROGRAM");
146 msg (SE, _("The %s subcommand may only be specified once."), "END");
150 lex_match (lexer, T_EQUALS);
151 if (!lex_force_id (lexer))
153 end = dict_lookup_var (dict, lex_tokcstr (lexer));
155 end = dict_create_var_assert (dict, lex_tokcstr (lexer), 0);
158 else if (lex_match_id (lexer, "NOTABLE"))
160 else if (lex_match_id (lexer, "TABLE"))
162 else if (lex_token (lexer) == T_ID)
164 if (lex_match_id (lexer, "FIXED"))
165 data_parser_set_type (parser, DP_FIXED);
166 else if (lex_match_id (lexer, "FREE"))
168 data_parser_set_type (parser, DP_DELIMITED);
169 data_parser_set_span (parser, true);
171 else if (lex_match_id (lexer, "LIST"))
173 data_parser_set_type (parser, DP_DELIMITED);
174 data_parser_set_span (parser, false);
178 lex_error (lexer, NULL);
184 msg (SE, _("Only one of FIXED, FREE, or LIST may "
190 if (data_parser_get_type (parser) == DP_DELIMITED)
192 if (lex_match (lexer, T_LPAREN))
194 struct string delims = DS_EMPTY_INITIALIZER;
196 while (!lex_match (lexer, T_RPAREN))
200 if (lex_match_id (lexer, "TAB"))
202 else if (lex_is_string (lexer)
203 && ss_length (lex_tokss (lexer)) == 1)
205 delim = ss_first (lex_tokss (lexer));
210 /* XXX should support multibyte UTF-8 characters */
211 lex_error (lexer, NULL);
212 ds_destroy (&delims);
215 ds_put_byte (&delims, delim);
217 lex_match (lexer, T_COMMA);
220 data_parser_set_empty_line_has_field (parser, true);
221 data_parser_set_quotes (parser, ss_empty ());
222 data_parser_set_soft_delimiters (parser, ss_empty ());
223 data_parser_set_hard_delimiters (parser, ds_ss (&delims));
224 ds_destroy (&delims);
228 data_parser_set_empty_line_has_field (parser, false);
229 data_parser_set_quotes (parser, ss_cstr ("'\""));
230 data_parser_set_soft_delimiters (parser,
231 ss_cstr (CC_SPACES));
232 const char decimal = settings_get_decimal_char (FMT_F);
233 data_parser_set_hard_delimiters (parser,
234 ss_buffer (",", (decimal == '.') ? 1 : 0));
240 lex_error (lexer, NULL);
244 type = data_parser_get_type (parser);
246 if (encoding && NULL == fh)
247 msg (MW, _("Encoding should not be specified for inline data. It will be "
251 fh = fh_inline_file ();
252 fh_set_default_handle (fh);
254 if (type != DP_FIXED && end != NULL)
256 msg (SE, _("The %s subcommand may be used only with %s."), "END", "DATA LIST FIXED");
260 tmp_pool = pool_create ();
261 if (type == DP_FIXED)
262 ok = parse_fixed (lexer, dict, tmp_pool, parser);
264 ok = parse_free (lexer, dict, tmp_pool, parser);
265 pool_destroy (tmp_pool);
269 if (!data_parser_any_fields (parser))
271 msg (SE, _("At least one variable must be specified."));
275 if (lex_end_of_command (lexer) != CMD_SUCCESS)
279 table = type == DP_FIXED || !data_parser_get_span (parser);
281 data_parser_output_description (parser, fh);
283 reader = dfm_open_reader (fh, lexer, encoding);
287 if (in_input_program ())
289 struct data_list_trns *trns = xmalloc (sizeof *trns);
290 trns->parser = parser;
291 trns->reader = reader;
293 add_transformation (ds, data_list_trns_proc, data_list_trns_free, trns);
296 data_parser_make_active_file (parser, dataset_session (ds), reader, dict);
301 return CMD_DATA_LIST;
304 data_parser_destroy (parser);
305 if (!in_input_program ())
309 return CMD_CASCADING_FAILURE;
312 /* Fixed-format parsing. */
314 /* Parses all the variable specifications for DATA LIST FIXED,
315 storing them into DLS. Uses TMP_POOL for temporary storage;
316 the caller may destroy it. Returns true only if
319 parse_fixed (struct lexer *lexer, struct dictionary *dict,
320 struct pool *tmp_pool, struct data_parser *parser)
322 int max_records = data_parser_get_records (parser);
326 while (lex_token (lexer) != T_ENDCMD)
329 size_t name_cnt, name_idx;
330 struct fmt_spec *formats, *f;
333 /* Parse everything. */
334 if (!parse_record_placement (lexer, &record, &column)
335 || !parse_DATA_LIST_vars_pool (lexer, dict, tmp_pool,
336 &names, &name_cnt, PV_NONE)
337 || !parse_var_placements (lexer, tmp_pool, name_cnt, FMT_FOR_INPUT,
338 &formats, &format_cnt))
341 /* Create variables and var specs. */
343 for (f = formats; f < &formats[format_cnt]; f++)
344 if (!execute_placement_format (f, &record, &column))
350 name = names[name_idx++];
352 /* Create variable. */
353 width = fmt_var_width (f);
354 v = dict_create_var (dict, name, width);
358 struct fmt_spec output = fmt_for_output_from_input (f);
359 var_set_both_formats (v, &output);
364 This can be acceptable if we're in INPUT
365 PROGRAM, but only if the existing variable has
366 the same width as the one we would have
368 if (!in_input_program ())
370 msg (SE, _("%s is a duplicate variable name."), name);
374 v = dict_lookup_var_assert (dict, name);
375 if ((width != 0) != (var_get_width (v) != 0))
377 msg (SE, _("There is already a variable %s of a "
382 if (width != 0 && width != var_get_width (v))
384 msg (SE, _("There is already a string variable %s of a "
385 "different width."), name);
390 if (max_records && record > max_records)
392 msg (SE, _("Cannot place variable %s on record %d when "
393 "RECORDS=%d is specified."),
394 var_get_name (v), record,
395 data_parser_get_records (parser));
398 data_parser_add_fixed_field (parser, f,
399 var_get_case_index (v),
400 var_get_name (v), record, column);
404 assert (name_idx == name_cnt);
410 /* Free-format parsing. */
412 /* Parses variable specifications for DATA LIST FREE and adds
413 them to DLS. Uses TMP_POOL for temporary storage; the caller
414 may destroy it. Returns true only if successful. */
416 parse_free (struct lexer *lexer, struct dictionary *dict,
417 struct pool *tmp_pool, struct data_parser *parser)
420 while (lex_token (lexer) != T_ENDCMD)
422 struct fmt_spec input, output;
427 if (!parse_DATA_LIST_vars_pool (lexer, dict, tmp_pool,
428 &name, &name_cnt, PV_NONE))
431 if (lex_match (lexer, T_LPAREN))
433 char type[FMT_TYPE_LEN_MAX + 1];
435 if (!parse_abstract_format_specifier (lexer, type, &input.w,
438 if (!fmt_from_name (type, &input.type))
440 msg (SE, _("Unknown format type `%s'."), type);
444 /* If no width was included, use the minimum width for the type.
445 This isn't quite right, because DATETIME by itself seems to become
446 DATETIME20 (see bug #30690), whereas this will become
447 DATETIME17. The correct behavior is not documented. */
450 input.w = fmt_min_input_width (input.type);
454 if (!fmt_check_input (&input) || !lex_force_match (lexer, T_RPAREN))
457 /* As a special case, N format is treated as F format
458 for free-field input. */
459 if (input.type == FMT_N)
462 output = fmt_for_output_from_input (&input);
466 lex_match (lexer, T_ASTERISK);
467 input = fmt_for_input (FMT_F, 8, 0);
468 output = *settings_get_format ();
471 for (i = 0; i < name_cnt; i++)
475 v = dict_create_var (dict, name[i], fmt_var_width (&input));
478 msg (SE, _("%s is a duplicate variable name."), name[i]);
481 var_set_both_formats (v, &output);
483 data_parser_add_delimited_field (parser,
484 &input, var_get_case_index (v),
492 /* Input procedure. */
494 /* Destroys DATA LIST transformation TRNS.
495 Returns true if successful, false if an I/O error occurred. */
497 data_list_trns_free (void *trns_)
499 struct data_list_trns *trns = trns_;
500 data_parser_destroy (trns->parser);
501 dfm_close_reader (trns->reader);
506 /* Handle DATA LIST transformation TRNS, parsing data into *C. */
508 data_list_trns_proc (void *trns_, struct ccase **c, casenumber case_num UNUSED)
510 struct data_list_trns *trns = trns_;
513 *c = case_unshare (*c);
514 if (data_parser_parse (trns->parser, trns->reader, *c))
515 retval = TRNS_CONTINUE;
516 else if (dfm_reader_error (trns->reader) || dfm_eof (trns->reader) > 1)
518 /* An I/O error, or encountering end of file for a second
519 time, should be escalated into a more serious error. */
523 retval = TRNS_END_FILE;
525 /* If there was an END subcommand handle it. */
526 if (trns->end != NULL)
528 double *end = &case_data_rw (*c, trns->end)->f;
529 if (retval == TRNS_END_FILE)
532 retval = TRNS_CONTINUE;