1 /* PSPP - a program for statistical analysis.
2 Copyright (C) 1997-9, 2000, 2006, 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/>. */
21 #include "data/case.h"
22 #include "data/dataset.h"
23 #include "data/data-out.h"
24 #include "data/format.h"
25 #include "data/transformations.h"
26 #include "data/variable.h"
27 #include "language/command.h"
28 #include "language/data-io/data-writer.h"
29 #include "language/data-io/file-handle.h"
30 #include "language/data-io/placement-parser.h"
31 #include "language/lexer/format-parser.h"
32 #include "language/lexer/lexer.h"
33 #include "language/lexer/variable-parser.h"
34 #include "libpspp/assertion.h"
35 #include "libpspp/compiler.h"
36 #include "libpspp/i18n.h"
37 #include "libpspp/ll.h"
38 #include "libpspp/message.h"
39 #include "libpspp/misc.h"
40 #include "libpspp/pool.h"
41 #include "output/tab.h"
42 #include "output/text-item.h"
44 #include "gl/xalloc.h"
47 #define _(msgid) gettext (msgid)
49 /* Describes what to do when an output field is encountered. */
52 PRT_LITERAL, /* Literal string. */
53 PRT_VAR /* Variable. */
56 /* Describes how to output one field. */
60 struct ll ll; /* In struct print_trns `specs' list. */
61 enum field_type type; /* What type of field this is. */
62 int record; /* 1-based record number. */
63 int first_column; /* 0-based first column. */
66 const struct variable *var; /* Associated variable. */
67 struct fmt_spec format; /* Output spec. */
68 bool add_space; /* Add trailing space? */
69 bool sysmis_as_spaces; /* Output SYSMIS as spaces? */
71 /* PRT_LITERAL only. */
72 struct string string; /* String to output. */
75 static inline struct prt_out_spec *
76 ll_to_prt_out_spec (struct ll *ll)
78 return ll_data (ll, struct prt_out_spec, ll);
81 /* PRINT, PRINT EJECT, WRITE private data structure. */
84 struct pool *pool; /* Stores related data. */
85 bool eject; /* Eject page before printing? */
86 bool include_prefix; /* Prefix lines with space? */
87 const char *encoding; /* Encoding to use for output. */
88 struct dfm_writer *writer; /* Output file, NULL=listing file. */
89 struct ll_list specs; /* List of struct prt_out_specs. */
90 size_t record_cnt; /* Number of records to write. */
91 struct u8_line line; /* Output buffer. */
93 int unit; /* Unit width, in bytes. */
94 char one[MAX_UNIT]; /* '1' in encoding, 'unit' bytes long. */
95 char space[MAX_UNIT]; /* \n in encoding, 'unit' bytes long. */
104 static int internal_cmd_print (struct lexer *, struct dataset *ds,
105 enum which_formats, bool eject);
106 static trns_proc_func print_trns_proc;
107 static trns_free_func print_trns_free;
108 static bool parse_specs (struct lexer *, struct pool *tmp_pool, struct print_trns *,
109 struct dictionary *dict, enum which_formats);
110 static void dump_table (struct print_trns *, const struct file_handle *);
114 /* Parses PRINT command. */
116 cmd_print (struct lexer *lexer, struct dataset *ds)
118 return internal_cmd_print (lexer, ds, PRINT, false);
121 /* Parses PRINT EJECT command. */
123 cmd_print_eject (struct lexer *lexer, struct dataset *ds)
125 return internal_cmd_print (lexer, ds, PRINT, true);
128 /* Parses WRITE command. */
130 cmd_write (struct lexer *lexer, struct dataset *ds)
132 return internal_cmd_print (lexer, ds, WRITE, false);
135 /* Parses the output commands. */
137 internal_cmd_print (struct lexer *lexer, struct dataset *ds,
138 enum which_formats which_formats, bool eject)
140 bool print_table = 0;
141 struct print_trns *trns;
142 struct file_handle *fh = NULL;
143 char *encoding = NULL;
144 struct pool *tmp_pool;
146 /* Fill in prt to facilitate error-handling. */
147 trns = pool_create_container (struct print_trns, pool);
150 trns->record_cnt = 0;
151 ll_init (&trns->specs);
152 u8_line_init (&trns->line);
153 u8_line_register_pool (&trns->line, trns->pool);
155 tmp_pool = pool_create_subpool (trns->pool);
157 /* Parse the command options. */
158 while (lex_token (lexer) != T_SLASH && lex_token (lexer) != T_ENDCMD)
160 if (lex_match_id (lexer, "OUTFILE"))
162 lex_match (lexer, T_EQUALS);
164 fh = fh_parse (lexer, FH_REF_FILE, NULL);
168 else if (lex_match_id (lexer, "ENCODING"))
170 lex_match (lexer, T_EQUALS);
171 if (!lex_force_string (lexer))
175 encoding = ss_xstrdup (lex_tokss (lexer));
179 else if (lex_match_id (lexer, "RECORDS"))
181 lex_match (lexer, T_EQUALS);
182 lex_match (lexer, T_LPAREN);
183 if (!lex_force_int (lexer))
185 trns->record_cnt = lex_integer (lexer);
187 lex_match (lexer, T_RPAREN);
189 else if (lex_match_id (lexer, "TABLE"))
191 else if (lex_match_id (lexer, "NOTABLE"))
195 lex_error (lexer, _("expecting a valid subcommand"));
200 /* When PRINT or PRINT EJECT writes to an external file, we
201 prefix each line with a space for compatibility. */
202 trns->include_prefix = which_formats == PRINT && fh != NULL;
204 /* Parse variables and strings. */
205 if (!parse_specs (lexer, tmp_pool, trns, dataset_dict (ds), which_formats))
208 if (lex_end_of_command (lexer) != CMD_SUCCESS)
213 trns->writer = dfm_open_writer (fh, encoding);
214 if (trns->writer == NULL)
216 trns->encoding = dfm_writer_get_encoding (trns->writer);
219 trns->encoding = UTF8;
221 /* Output the variable table if requested. */
223 dump_table (trns, fh);
225 /* Put the transformation in the queue. */
226 add_transformation (ds, print_trns_proc, print_trns_free, trns);
228 pool_destroy (tmp_pool);
234 print_trns_free (trns);
239 static bool parse_string_argument (struct lexer *, struct print_trns *,
240 int record, int *column);
241 static bool parse_variable_argument (struct lexer *, const struct dictionary *,
243 struct pool *tmp_pool,
244 int *record, int *column,
247 /* Parses all the variable and string specifications on a single
248 PRINT, PRINT EJECT, or WRITE command into the prt structure.
251 parse_specs (struct lexer *lexer, struct pool *tmp_pool, struct print_trns *trns,
252 struct dictionary *dict,
253 enum which_formats which_formats)
258 if (lex_token (lexer) == T_ENDCMD)
260 trns->record_cnt = 1;
264 while (lex_token (lexer) != T_ENDCMD)
268 if (!parse_record_placement (lexer, &record, &column))
271 if (lex_is_string (lexer))
272 ok = parse_string_argument (lexer, trns, record, &column);
274 ok = parse_variable_argument (lexer, dict, trns, tmp_pool, &record, &column,
279 lex_match (lexer, T_COMMA);
282 if (trns->record_cnt != 0 && trns->record_cnt != record)
283 msg (SW, _("Output calls for %d records but %zu specified on RECORDS "
285 record, trns->record_cnt);
286 trns->record_cnt = record;
291 /* Parses a string argument to the PRINT commands. Returns success. */
293 parse_string_argument (struct lexer *lexer, struct print_trns *trns, int record, int *column)
295 struct prt_out_spec *spec = pool_alloc (trns->pool, sizeof *spec);
296 spec->type = PRT_LITERAL;
297 spec->record = record;
298 spec->first_column = *column;
299 ds_init_substring (&spec->string, lex_tokss (lexer));
300 ds_register_pool (&spec->string, trns->pool);
303 /* Parse the included column range. */
304 if (lex_is_number (lexer))
306 int first_column, last_column;
307 bool range_specified;
309 if (!parse_column_range (lexer, 1,
310 &first_column, &last_column, &range_specified))
313 spec->first_column = first_column;
315 ds_set_length (&spec->string, last_column - first_column + 1, ' ');
317 *column = spec->first_column + ds_length (&spec->string);
319 ll_push_tail (&trns->specs, &spec->ll);
323 /* Parses a variable argument to the PRINT commands by passing it off
324 to fixed_parse_compatible() or fixed_parse_fortran() as appropriate.
327 parse_variable_argument (struct lexer *lexer, const struct dictionary *dict,
328 struct print_trns *trns, struct pool *tmp_pool,
329 int *record, int *column,
330 enum which_formats which_formats)
332 const struct variable **vars;
333 size_t var_cnt, var_idx;
334 struct fmt_spec *formats, *f;
338 if (!parse_variables_const_pool (lexer, tmp_pool, dict,
339 &vars, &var_cnt, PV_DUPLICATE))
342 if (lex_is_number (lexer) || lex_token (lexer) == T_LPAREN)
344 if (!parse_var_placements (lexer, tmp_pool, var_cnt, FMT_FOR_OUTPUT,
345 &formats, &format_cnt))
353 lex_match (lexer, T_ASTERISK);
355 formats = pool_nmalloc (tmp_pool, var_cnt, sizeof *formats);
356 format_cnt = var_cnt;
357 for (i = 0; i < var_cnt; i++)
359 const struct variable *v = vars[i];
360 formats[i] = (which_formats == PRINT
361 ? *var_get_print_format (v)
362 : *var_get_write_format (v));
364 add_space = which_formats == PRINT;
368 for (f = formats; f < &formats[format_cnt]; f++)
369 if (!execute_placement_format (f, record, column))
371 const struct variable *var;
372 struct prt_out_spec *spec;
374 var = vars[var_idx++];
375 if (!fmt_check_width_compat (f, var_get_width (var)))
378 spec = pool_alloc (trns->pool, sizeof *spec);
379 spec->type = PRT_VAR;
380 spec->record = *record;
381 spec->first_column = *column;
384 spec->add_space = add_space;
386 /* This is a completely bizarre twist for compatibility:
387 WRITE outputs the system-missing value as a field
388 filled with spaces, instead of using the normal format
389 that usually contains a period. */
390 spec->sysmis_as_spaces = (which_formats == WRITE
391 && var_is_numeric (var)
392 && (fmt_get_category (spec->format.type)
395 ll_push_tail (&trns->specs, &spec->ll);
397 *column += f->w + add_space;
399 assert (var_idx == var_cnt);
404 /* Prints the table produced by the TABLE subcommand to the listing
407 dump_table (struct print_trns *trns, const struct file_handle *fh)
409 struct prt_out_spec *spec;
414 spec_cnt = ll_count (&trns->specs);
415 t = tab_create (4, spec_cnt + 1);
416 tab_box (t, TAL_1, TAL_1, TAL_0, TAL_1, 0, 0, 3, spec_cnt);
417 tab_hline (t, TAL_2, 0, 3, 1);
418 tab_headers (t, 0, 0, 1, 0);
419 tab_text (t, 0, 0, TAB_CENTER | TAT_TITLE, _("Variable"));
420 tab_text (t, 1, 0, TAB_CENTER | TAT_TITLE, _("Record"));
421 tab_text (t, 2, 0, TAB_CENTER | TAT_TITLE, _("Columns"));
422 tab_text (t, 3, 0, TAB_CENTER | TAT_TITLE, _("Format"));
424 ll_for_each (spec, struct prt_out_spec, ll, &trns->specs)
426 char fmt_string[FMT_STRING_LEN_MAX + 1];
431 tab_text_format (t, 0, row, TAB_LEFT | TAB_FIX, "`%.*s'",
432 (int) ds_length (&spec->string),
433 ds_data (&spec->string));
434 width = ds_length (&spec->string);
437 tab_text (t, 0, row, TAB_LEFT, var_get_name (spec->var));
438 tab_text (t, 3, row, TAB_LEFT | TAB_FIX,
439 fmt_to_string (&spec->format, fmt_string));
440 width = spec->format.w;
445 tab_text_format (t, 1, row, 0, "%d", spec->record);
446 tab_text_format (t, 2, row, 0, "%3d-%3d",
447 spec->first_column, spec->first_column + width - 1);
452 tab_title (t, ngettext ("Writing %zu record to %s.",
453 "Writing %zu records to %s.", trns->record_cnt),
454 trns->record_cnt, fh_get_name (fh));
456 tab_title (t, ngettext ("Writing %zu record.",
457 "Writing %zu records.", trns->record_cnt),
462 /* Transformation. */
464 static void flush_records (struct print_trns *, int target_record,
465 bool *eject, int *record);
467 /* Performs the transformation inside print_trns T on case C. */
469 print_trns_proc (void *trns_, struct ccase **c, casenumber case_num UNUSED)
471 struct print_trns *trns = trns_;
472 bool eject = trns->eject;
473 char encoded_space = recode_byte (trns->encoding, C_ENCODING, ' ');
475 struct prt_out_spec *spec;
477 u8_line_clear (&trns->line);
479 ds_put_byte (&trns->line.s, ' ');
480 trns->line.width = 0;
482 ll_for_each (spec, struct prt_out_spec, ll, &trns->specs)
484 flush_records (trns, spec->record, &eject, &record);
486 if (spec->type == PRT_VAR)
488 const union value *input = case_data (*c, spec->var);
489 if (!spec->sysmis_as_spaces || input->f != SYSMIS)
491 char *s = data_out (input, var_get_encoding (spec->var),
493 int width = u8_strwidth (s);
494 size_t n = strlen (s);
495 u8_line_put (&trns->line, spec->first_column,
496 spec->first_column + width, s, n);
501 memset (u8_line-record
504 ds_put_byte_multiple (&trns->line, encoded_space, spec->format.w);
506 ds_put_byte (&trns->line, encoded_space);
510 ds_put_substring (&trns->line, ds_ss (&spec->string));
511 if (0 != strcmp (trns->encoding, UTF8))
513 size_t length = ds_length (&spec->string);
514 char *data = ss_data (ds_tail (&trns->line, length));
515 char *s = recode_string (trns->encoding, UTF8, data, length);
516 memcpy (data, s, length);
521 flush_records (trns, trns->record_cnt + 1, &eject, &record);
523 if (trns->writer != NULL && dfm_write_error (trns->writer))
525 return TRNS_CONTINUE;
528 /* Advance from *RECORD to TARGET_RECORD, outputting records
529 along the way. If *EJECT is true, then the first record
530 output is preceded by ejecting the page (and *EJECT is set
533 flush_records (struct print_trns *trns, int target_record,
534 bool *eject, int *record)
536 for (; target_record > *record; (*record)++)
538 char *line = ds_cstr (&trns->line);
539 size_t length = ds_length (&trns->line);
545 if (trns->writer == NULL)
546 text_item_submit (text_item_create (TEXT_ITEM_EJECT_PAGE, ""));
550 line[0] = recode_byte (trns->encoding, C_ENCODING, leader);
552 if (trns->writer == NULL)
553 tab_output_text (TAB_FIX, &line[1]);
556 if (!trns->include_prefix)
561 dfm_put_record (trns->writer, line, length);
564 ds_truncate (&trns->line, 1);
570 print_trns_free (void *trns_)
572 struct print_trns *trns = trns_;
575 if (trns->writer != NULL)
576 ok = dfm_close_writer (trns->writer);
577 pool_destroy (trns->pool);