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 <libpspp/alloc.h>
42 #include <libpspp/array.h>
43 #include <libpspp/message.h>
44 #include <libpspp/message.h>
45 #include <libpspp/misc.h>
46 #include <libpspp/pool.h>
47 #include <libpspp/str.h>
52 #define _(msgid) gettext (msgid)
54 /* List of variable names. */
58 char name[SHORT_NAME_LEN + 1];
61 /* Represents a FLIP input program. */
64 struct pool *pool; /* Pool containing FLIP data. */
65 struct variable **var; /* Variables to transpose. */
66 int *idx_to_fv; /* var[]->index to compacted sink case fv. */
67 size_t var_cnt; /* Number of elements in `var'. */
68 int case_cnt; /* Pre-flip case count. */
69 size_t case_size; /* Post-flip bytes per case. */
71 union value *output_buf; /* Case output buffer. */
73 struct variable *new_names; /* Variable containing new variable names. */
74 struct varname *new_names_head; /* First new variable. */
75 struct varname *new_names_tail; /* Last new variable. */
77 FILE *file; /* Temporary file containing data. */
80 static void destroy_flip_pgm (struct flip_pgm *);
81 static struct case_sink *flip_sink_create (struct flip_pgm *);
82 static struct case_source *flip_source_create (struct flip_pgm *);
83 static bool flip_file (struct flip_pgm *);
84 static int build_dictionary (struct flip_pgm *);
86 static const struct case_source_class flip_source_class;
87 static const struct case_sink_class flip_sink_class;
89 /* Parses and executes FLIP. */
93 struct flip_pgm *flip;
94 struct case_sink *sink;
97 if (proc_make_temporary_transformations_permanent ())
98 msg (SW, _("FLIP ignores TEMPORARY. "
99 "Temporary transformations will be made permanent."));
101 flip = pool_create_container (struct flip_pgm, pool);
103 flip->idx_to_fv = dict_get_compacted_idx_to_fv (default_dict);
104 pool_register (flip->pool, free, flip->idx_to_fv);
107 flip->new_names = NULL;
108 flip->new_names_head = NULL;
109 flip->new_names_tail = NULL;
113 if (lex_match_id ("VARIABLES"))
116 if (!parse_variables (default_dict, &flip->var, &flip->var_cnt,
122 dict_get_vars (default_dict, &flip->var, &flip->var_cnt, 1u << DC_SYSTEM);
123 pool_register (flip->pool, free, flip->var);
126 if (lex_match_id ("NEWNAMES"))
129 flip->new_names = parse_variable ();
130 if (!flip->new_names)
134 flip->new_names = dict_lookup_var (default_dict, "CASE_LBL");
140 for (i = 0; i < flip->var_cnt; i++)
141 if (flip->var[i] == flip->new_names)
143 remove_element (flip->var, flip->var_cnt, sizeof *flip->var, i);
149 /* Read the active file into a flip_sink. */
151 proc_make_temporary_transformations_permanent ();
152 sink = flip_sink_create (flip);
155 proc_set_sink (sink);
156 flip->new_names_tail = NULL;
157 ok = procedure (NULL, NULL);
159 /* Flip the data we read. */
160 if (!flip_file (flip))
162 discard_variables ();
166 /* Flip the dictionary. */
167 dict_clear (default_dict);
168 if (!build_dictionary (flip))
170 discard_variables ();
173 flip->case_size = dict_get_case_size (default_dict);
175 /* Set up flipped data for reading. */
176 proc_set_source (flip_source_create (flip));
178 return ok ? lex_end_of_command () : CMD_CASCADING_FAILURE;
181 destroy_flip_pgm (flip);
182 return CMD_CASCADING_FAILURE;
187 destroy_flip_pgm (struct flip_pgm *flip)
190 pool_destroy (flip->pool);
193 /* Make a new variable with base name NAME, which is bowdlerized and
194 mangled until acceptable, and returns success. */
196 make_new_var (char name[])
200 /* Trim trailing spaces. */
201 cp = strchr (name, '\0');
202 while (cp > name && isspace ((unsigned char) cp[-1]))
205 /* Fix invalid characters. */
206 for (cp = name; *cp && cp < name + SHORT_NAME_LEN; cp++)
209 if (!lex_is_id1 (*cp) || *cp == '$')
214 if (!lex_is_idn (*cp))
218 str_uppercase (name);
220 if (dict_create_var (default_dict, name, 0))
223 /* Add numeric extensions until acceptable. */
225 const int len = (int) strlen (name);
226 char n[SHORT_NAME_LEN + 1];
229 for (i = 1; i < 10000000; i++)
231 int ofs = min (7 - intlog10 (i), len);
232 memcpy (n, name, ofs);
233 sprintf (&n[ofs], "%d", i);
235 if (dict_create_var (default_dict, n, 0))
240 msg (SE, _("Could not create acceptable variant for variable %s."), name);
244 /* Make a new dictionary for all the new variable names. */
246 build_dictionary (struct flip_pgm *flip)
248 dict_create_var_assert (default_dict, "CASE_LBL", 8);
250 if (flip->new_names_head == NULL)
254 if (flip->case_cnt > 99999)
256 msg (SE, _("Cannot create more than 99999 variable names."));
260 for (i = 0; i < flip->case_cnt; i++)
263 char s[SHORT_NAME_LEN + 1];
265 sprintf (s, "VAR%03d", i);
266 v = dict_create_var_assert (default_dict, s, 0);
273 for (v = flip->new_names_head; v; v = v->next)
274 if (!make_new_var (v->name))
281 /* Creates a flip sink based on FLIP. */
282 static struct case_sink *
283 flip_sink_create (struct flip_pgm *flip)
287 flip->output_buf = pool_nalloc (flip->pool,
288 flip->var_cnt, sizeof *flip->output_buf);
290 flip->file = pool_tmpfile (flip->pool);
291 if (flip->file == NULL)
293 msg (SE, _("Could not create temporary file for FLIP."));
297 /* Write variable names as first case. */
298 for (i = 0; i < flip->var_cnt; i++)
299 buf_copy_str_rpad (flip->output_buf[i].s, MAX_SHORT_STRING,
301 if (fwrite (flip->output_buf, sizeof *flip->output_buf,
302 flip->var_cnt, flip->file) != (size_t) flip->var_cnt)
304 msg (SE, _("Error writing FLIP file: %s."), strerror (errno));
310 return create_case_sink (&flip_sink_class, default_dict, flip);
313 /* Writes case C to the FLIP sink.
314 Returns true if successful, false if an I/O error occurred. */
316 flip_sink_write (struct case_sink *sink, const struct ccase *c)
318 struct flip_pgm *flip = sink->aux;
323 if (flip->new_names != NULL)
325 struct varname *v = pool_alloc (flip->pool, sizeof *v);
327 if (flip->new_names->type == NUMERIC)
329 double f = case_num (c, flip->idx_to_fv[flip->new_names->index]);
332 strcpy (v->name, "VSYSMIS");
333 else if (f < INT_MIN)
334 strcpy (v->name, "VNEGINF");
335 else if (f > INT_MAX)
336 strcpy (v->name, "VPOSINF");
338 snprintf (v->name, sizeof v->name, "V%d", (int) f);
342 int width = min (flip->new_names->width, MAX_SHORT_STRING);
343 memcpy (v->name, case_str (c, flip->idx_to_fv[flip->new_names->index]),
348 if (flip->new_names_head == NULL)
349 flip->new_names_head = v;
351 flip->new_names_tail->next = v;
352 flip->new_names_tail = v;
355 /* Write to external file. */
356 for (i = 0; i < flip->var_cnt; i++)
360 if (flip->var[i]->type == NUMERIC)
361 out = case_num (c, flip->idx_to_fv[flip->var[i]->index]);
364 flip->output_buf[i].f = out;
367 if (fwrite (flip->output_buf, sizeof *flip->output_buf,
368 flip->var_cnt, flip->file) != (size_t) flip->var_cnt)
370 msg (SE, _("Error writing FLIP file: %s."), strerror (errno));
376 /* Transposes the external file into a new file. */
378 flip_file (struct flip_pgm *flip)
381 size_t case_capacity;
383 union value *input_buf, *output_buf;
384 FILE *input_file, *output_file;
386 /* Allocate memory for many cases. */
387 case_bytes = flip->var_cnt * sizeof *input_buf;
388 case_capacity = get_workspace () / case_bytes;
389 if (case_capacity > flip->case_cnt * 2)
390 case_capacity = flip->case_cnt * 2;
391 if (case_capacity < 2)
395 size_t bytes = case_bytes * case_capacity;
396 if (case_capacity > 2)
397 input_buf = malloc (bytes);
399 input_buf = xmalloc (bytes);
400 if (input_buf != NULL)
404 if (case_capacity < 2)
407 pool_register (flip->pool, free, input_buf);
409 /* Use half the allocated memory for input_buf, half for
412 output_buf = input_buf + flip->var_cnt * case_capacity;
414 input_file = flip->file;
415 if (fseek (input_file, 0, SEEK_SET) != 0)
417 msg (SE, _("Error rewinding FLIP file: %s."), strerror (errno));
421 output_file = pool_tmpfile (flip->pool);
422 if (output_file == NULL)
424 msg (SE, _("Error creating FLIP source file."));
428 for (case_idx = 0; case_idx < flip->case_cnt; )
430 unsigned long read_cases = min (flip->case_cnt - case_idx,
434 if (read_cases != fread (input_buf, case_bytes, read_cases, input_file))
436 msg (SE, _("Error reading FLIP file: %s."), strerror (errno));
440 for (i = 0; i < flip->var_cnt; i++)
444 for (j = 0; j < read_cases; j++)
445 output_buf[j] = input_buf[i + j * flip->var_cnt];
452 #define off_t long int
455 if (fseeko (output_file,
456 sizeof *input_buf * (case_idx
457 + (off_t) i * flip->case_cnt),
460 msg (SE, _("Error seeking FLIP source file: %s."),
465 if (fwrite (output_buf, sizeof *output_buf, read_cases, output_file)
468 msg (SE, _("Error writing FLIP source file: %s."),
474 case_idx += read_cases;
477 if (pool_fclose (flip->pool, input_file) == EOF)
479 msg (SE, _("Error closing FLIP source file: %s."), strerror (errno));
482 pool_unregister (flip->pool, input_buf);
485 if (fseek (output_file, 0, SEEK_SET) != 0)
487 msg (SE, _("Error rewinding FLIP source file: %s."), strerror (errno));
490 flip->file = output_file;
495 /* FLIP sink class. */
496 static const struct case_sink_class flip_sink_class =
505 /* Creates and returns a FLIP source based on PGM,
506 which should have already been used as a sink. */
507 static struct case_source *
508 flip_source_create (struct flip_pgm *pgm)
510 return create_case_source (&flip_source_class, pgm);
513 /* Reads the FLIP stream. Copies each case into C and calls
514 WRITE_CASE passing WC_DATA.
515 Returns true if successful, false if an I/O error occurred. */
517 flip_source_read (struct case_source *source,
519 write_case_func *write_case, write_case_data wc_data)
521 struct flip_pgm *flip = source->aux;
522 union value *input_buf;
526 input_buf = xnmalloc (flip->case_cnt, sizeof *input_buf);
527 for (i = 0; ok && i < flip->var_cnt; i++)
531 if (fread (input_buf, sizeof *input_buf, flip->case_cnt,
532 flip->file) != flip->case_cnt)
534 if (ferror (flip->file))
535 msg (SE, _("Error reading FLIP temporary file: %s."),
537 else if (feof (flip->file))
538 msg (SE, _("Unexpected end of file reading FLIP temporary file."));
545 for (j = 0; j < flip->case_cnt; j++)
546 case_data_rw (c, j)->f = input_buf[j].f;
547 ok = write_case (wc_data);
554 /* Destroy internal data in SOURCE. */
556 flip_source_destroy (struct case_source *source)
558 struct flip_pgm *flip = source->aux;
560 destroy_flip_pgm (flip);
563 static const struct case_source_class flip_source_class =