1 /* PSPP - computes sample statistics.
2 Copyright (C) 1997-9, 2000 Free Software Foundation, Inc.
3 Written by Ben Pfaff <blp@gnu.org>.
5 This program is free software; you can redistribute it and/or
6 modify it under the terms of the GNU General Public License as
7 published by the Free Software Foundation; either version 2 of the
8 License, or (at your option) any later version.
10 This program is distributed in the hope that it will be useful, but
11 WITHOUT ANY WARRANTY; without even the implied warranty of
12 MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
13 General Public License for more details.
15 You should have received a copy of the GNU General Public License
16 along with this program; if not, write to the Free Software
17 Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA
27 #ifdef HAVE_SYS_TYPES_H
28 #include <sys/types.h>
31 #include <data/case-sink.h>
32 #include <data/case-source.h>
33 #include <data/case.h>
34 #include <data/dictionary.h>
35 #include <data/procedure.h>
36 #include <data/settings.h>
37 #include <data/value.h>
38 #include <data/variable.h>
39 #include <language/command.h>
40 #include <language/lexer/lexer.h>
41 #include <language/lexer/variable-parser.h>
42 #include <libpspp/alloc.h>
43 #include <libpspp/array.h>
44 #include <libpspp/message.h>
45 #include <libpspp/message.h>
46 #include <libpspp/misc.h>
47 #include <libpspp/pool.h>
48 #include <libpspp/str.h>
53 #define _(msgid) gettext (msgid)
55 /* List of variable names. */
59 char name[SHORT_NAME_LEN + 1];
62 /* Represents a FLIP input program. */
65 struct pool *pool; /* Pool containing FLIP data. */
66 struct variable **var; /* Variables to transpose. */
67 int *idx_to_fv; /* var[]->index to compacted sink case fv. */
68 size_t var_cnt; /* Number of elements in `var'. */
69 int case_cnt; /* Pre-flip case count. */
70 size_t case_size; /* Post-flip bytes per case. */
72 union value *output_buf; /* Case output buffer. */
74 struct variable *new_names; /* Variable containing new variable names. */
75 struct varname *new_names_head; /* First new variable. */
76 struct varname *new_names_tail; /* Last new variable. */
78 FILE *file; /* Temporary file containing data. */
81 static void destroy_flip_pgm (struct flip_pgm *);
82 static struct case_sink *flip_sink_create (struct flip_pgm *);
83 static struct case_source *flip_source_create (struct flip_pgm *);
84 static bool flip_file (struct flip_pgm *);
85 static int build_dictionary (struct flip_pgm *);
87 static const struct case_source_class flip_source_class;
88 static const struct case_sink_class flip_sink_class;
90 /* Parses and executes FLIP. */
94 struct flip_pgm *flip;
95 struct case_sink *sink;
98 if (proc_make_temporary_transformations_permanent ())
99 msg (SW, _("FLIP ignores TEMPORARY. "
100 "Temporary transformations will be made permanent."));
102 flip = pool_create_container (struct flip_pgm, pool);
104 flip->idx_to_fv = dict_get_compacted_idx_to_fv (default_dict);
105 pool_register (flip->pool, free, flip->idx_to_fv);
108 flip->new_names = NULL;
109 flip->new_names_head = NULL;
110 flip->new_names_tail = NULL;
114 if (lex_match_id ("VARIABLES"))
117 if (!parse_variables (default_dict, &flip->var, &flip->var_cnt,
123 dict_get_vars (default_dict, &flip->var, &flip->var_cnt, 1u << DC_SYSTEM);
124 pool_register (flip->pool, free, flip->var);
127 if (lex_match_id ("NEWNAMES"))
130 flip->new_names = parse_variable ();
131 if (!flip->new_names)
135 flip->new_names = dict_lookup_var (default_dict, "CASE_LBL");
141 for (i = 0; i < flip->var_cnt; i++)
142 if (flip->var[i] == flip->new_names)
144 remove_element (flip->var, flip->var_cnt, sizeof *flip->var, i);
150 /* Read the active file into a flip_sink. */
152 proc_make_temporary_transformations_permanent ();
153 sink = flip_sink_create (flip);
156 proc_set_sink (sink);
157 flip->new_names_tail = NULL;
158 ok = procedure (NULL, NULL);
160 /* Flip the data we read. */
161 if (!flip_file (flip))
163 discard_variables ();
167 /* Flip the dictionary. */
168 dict_clear (default_dict);
169 if (!build_dictionary (flip))
171 discard_variables ();
174 flip->case_size = dict_get_case_size (default_dict);
176 /* Set up flipped data for reading. */
177 proc_set_source (flip_source_create (flip));
179 return ok ? lex_end_of_command () : CMD_CASCADING_FAILURE;
182 destroy_flip_pgm (flip);
183 return CMD_CASCADING_FAILURE;
188 destroy_flip_pgm (struct flip_pgm *flip)
191 pool_destroy (flip->pool);
194 /* Make a new variable with base name NAME, which is bowdlerized and
195 mangled until acceptable, and returns success. */
197 make_new_var (char name[])
201 /* Trim trailing spaces. */
202 cp = strchr (name, '\0');
203 while (cp > name && isspace ((unsigned char) cp[-1]))
206 /* Fix invalid characters. */
207 for (cp = name; *cp && cp < name + SHORT_NAME_LEN; cp++)
210 if (!lex_is_id1 (*cp) || *cp == '$')
215 if (!lex_is_idn (*cp))
219 str_uppercase (name);
221 if (dict_create_var (default_dict, name, 0))
224 /* Add numeric extensions until acceptable. */
226 const int len = (int) strlen (name);
227 char n[SHORT_NAME_LEN + 1];
230 for (i = 1; i < 10000000; i++)
232 int ofs = min (7 - intlog10 (i), len);
233 memcpy (n, name, ofs);
234 sprintf (&n[ofs], "%d", i);
236 if (dict_create_var (default_dict, n, 0))
241 msg (SE, _("Could not create acceptable variant for variable %s."), name);
245 /* Make a new dictionary for all the new variable names. */
247 build_dictionary (struct flip_pgm *flip)
249 dict_create_var_assert (default_dict, "CASE_LBL", 8);
251 if (flip->new_names_head == NULL)
255 if (flip->case_cnt > 99999)
257 msg (SE, _("Cannot create more than 99999 variable names."));
261 for (i = 0; i < flip->case_cnt; i++)
264 char s[SHORT_NAME_LEN + 1];
266 sprintf (s, "VAR%03d", i);
267 v = dict_create_var_assert (default_dict, s, 0);
274 for (v = flip->new_names_head; v; v = v->next)
275 if (!make_new_var (v->name))
282 /* Creates a flip sink based on FLIP. */
283 static struct case_sink *
284 flip_sink_create (struct flip_pgm *flip)
288 flip->output_buf = pool_nalloc (flip->pool,
289 flip->var_cnt, sizeof *flip->output_buf);
291 flip->file = pool_tmpfile (flip->pool);
292 if (flip->file == NULL)
294 msg (SE, _("Could not create temporary file for FLIP."));
298 /* Write variable names as first case. */
299 for (i = 0; i < flip->var_cnt; i++)
300 buf_copy_str_rpad (flip->output_buf[i].s, MAX_SHORT_STRING,
302 if (fwrite (flip->output_buf, sizeof *flip->output_buf,
303 flip->var_cnt, flip->file) != (size_t) flip->var_cnt)
305 msg (SE, _("Error writing FLIP file: %s."), strerror (errno));
311 return create_case_sink (&flip_sink_class, default_dict, flip);
314 /* Writes case C to the FLIP sink.
315 Returns true if successful, false if an I/O error occurred. */
317 flip_sink_write (struct case_sink *sink, const struct ccase *c)
319 struct flip_pgm *flip = sink->aux;
324 if (flip->new_names != NULL)
326 struct varname *v = pool_alloc (flip->pool, sizeof *v);
328 if (flip->new_names->type == NUMERIC)
330 double f = case_num (c, flip->idx_to_fv[flip->new_names->index]);
333 strcpy (v->name, "VSYSMIS");
334 else if (f < INT_MIN)
335 strcpy (v->name, "VNEGINF");
336 else if (f > INT_MAX)
337 strcpy (v->name, "VPOSINF");
339 snprintf (v->name, sizeof v->name, "V%d", (int) f);
343 int width = min (flip->new_names->width, MAX_SHORT_STRING);
344 memcpy (v->name, case_str (c, flip->idx_to_fv[flip->new_names->index]),
349 if (flip->new_names_head == NULL)
350 flip->new_names_head = v;
352 flip->new_names_tail->next = v;
353 flip->new_names_tail = v;
356 /* Write to external file. */
357 for (i = 0; i < flip->var_cnt; i++)
361 if (flip->var[i]->type == NUMERIC)
362 out = case_num (c, flip->idx_to_fv[flip->var[i]->index]);
365 flip->output_buf[i].f = out;
368 if (fwrite (flip->output_buf, sizeof *flip->output_buf,
369 flip->var_cnt, flip->file) != (size_t) flip->var_cnt)
371 msg (SE, _("Error writing FLIP file: %s."), strerror (errno));
377 /* Transposes the external file into a new file. */
379 flip_file (struct flip_pgm *flip)
382 size_t case_capacity;
384 union value *input_buf, *output_buf;
385 FILE *input_file, *output_file;
387 /* Allocate memory for many cases. */
388 case_bytes = flip->var_cnt * sizeof *input_buf;
389 case_capacity = get_workspace () / case_bytes;
390 if (case_capacity > flip->case_cnt * 2)
391 case_capacity = flip->case_cnt * 2;
392 if (case_capacity < 2)
396 size_t bytes = case_bytes * case_capacity;
397 if (case_capacity > 2)
398 input_buf = malloc (bytes);
400 input_buf = xmalloc (bytes);
401 if (input_buf != NULL)
405 if (case_capacity < 2)
408 pool_register (flip->pool, free, input_buf);
410 /* Use half the allocated memory for input_buf, half for
413 output_buf = input_buf + flip->var_cnt * case_capacity;
415 input_file = flip->file;
416 if (fseek (input_file, 0, SEEK_SET) != 0)
418 msg (SE, _("Error rewinding FLIP file: %s."), strerror (errno));
422 output_file = pool_tmpfile (flip->pool);
423 if (output_file == NULL)
425 msg (SE, _("Error creating FLIP source file."));
429 for (case_idx = 0; case_idx < flip->case_cnt; )
431 unsigned long read_cases = min (flip->case_cnt - case_idx,
435 if (read_cases != fread (input_buf, case_bytes, read_cases, input_file))
437 msg (SE, _("Error reading FLIP file: %s."), strerror (errno));
441 for (i = 0; i < flip->var_cnt; i++)
445 for (j = 0; j < read_cases; j++)
446 output_buf[j] = input_buf[i + j * flip->var_cnt];
453 #define off_t long int
456 if (fseeko (output_file,
457 sizeof *input_buf * (case_idx
458 + (off_t) i * flip->case_cnt),
461 msg (SE, _("Error seeking FLIP source file: %s."),
466 if (fwrite (output_buf, sizeof *output_buf, read_cases, output_file)
469 msg (SE, _("Error writing FLIP source file: %s."),
475 case_idx += read_cases;
478 if (pool_fclose (flip->pool, input_file) == EOF)
480 msg (SE, _("Error closing FLIP source file: %s."), strerror (errno));
483 pool_unregister (flip->pool, input_buf);
486 if (fseek (output_file, 0, SEEK_SET) != 0)
488 msg (SE, _("Error rewinding FLIP source file: %s."), strerror (errno));
491 flip->file = output_file;
496 /* FLIP sink class. */
497 static const struct case_sink_class flip_sink_class =
506 /* Creates and returns a FLIP source based on PGM,
507 which should have already been used as a sink. */
508 static struct case_source *
509 flip_source_create (struct flip_pgm *pgm)
511 return create_case_source (&flip_source_class, pgm);
514 /* Reads the FLIP stream. Copies each case into C and calls
515 WRITE_CASE passing WC_DATA.
516 Returns true if successful, false if an I/O error occurred. */
518 flip_source_read (struct case_source *source,
520 write_case_func *write_case, write_case_data wc_data)
522 struct flip_pgm *flip = source->aux;
523 union value *input_buf;
527 input_buf = xnmalloc (flip->case_cnt, sizeof *input_buf);
528 for (i = 0; ok && i < flip->var_cnt; i++)
532 if (fread (input_buf, sizeof *input_buf, flip->case_cnt,
533 flip->file) != flip->case_cnt)
535 if (ferror (flip->file))
536 msg (SE, _("Error reading FLIP temporary file: %s."),
538 else if (feof (flip->file))
539 msg (SE, _("Unexpected end of file reading FLIP temporary file."));
546 for (j = 0; j < flip->case_cnt; j++)
547 case_data_rw (c, j)->f = input_buf[j].f;
548 ok = write_case (wc_data);
555 /* Destroy internal data in SOURCE. */
557 flip_source_destroy (struct case_source *source)
559 struct flip_pgm *flip = source->aux;
561 destroy_flip_pgm (flip);
564 static const struct case_source_class flip_source_class =