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., 59 Temple Place - Suite 330, Boston, MA
27 #include "algorithm.h"
31 #include "dictionary.h"
40 /* List of variable names. */
47 /* Represents a FLIP input program. */
50 struct variable **var; /* Variables to transpose. */
51 int *idx_to_fv; /* var[]->index to compacted sink case fv. */
52 int var_cnt; /* Number of elements in `var'. */
53 int case_cnt; /* Pre-flip case count. */
54 size_t case_size; /* Post-flip bytes per case. */
56 struct variable *new_names; /* Variable containing new variable names. */
57 struct varname *new_names_head; /* First new variable. */
58 struct varname *new_names_tail; /* Last new variable. */
60 FILE *file; /* Temporary file containing data. */
63 static void destroy_flip_pgm (struct flip_pgm *);
64 static struct case_sink *flip_sink_create (struct flip_pgm *);
65 static struct case_source *flip_source_create (struct flip_pgm *);
66 static void flip_file (struct flip_pgm *);
67 static int build_dictionary (struct flip_pgm *);
69 static const struct case_source_class flip_source_class;
70 static const struct case_sink_class flip_sink_class;
72 /* Parses and executes FLIP. */
76 struct flip_pgm *flip;
80 msg (SM, _("FLIP ignores TEMPORARY. "
81 "Temporary transformations will be made permanent."));
85 flip = xmalloc (sizeof *flip);
87 flip->idx_to_fv = dict_get_compacted_idx_to_fv (default_dict);
90 flip->new_names = NULL;
91 flip->new_names_head = NULL;
92 flip->new_names_tail = NULL;
96 if (lex_match_id ("VARIABLES"))
99 if (!parse_variables (default_dict, &flip->var, &flip->var_cnt, PV_NO_DUPLICATE))
104 dict_get_vars (default_dict, &flip->var, &flip->var_cnt, 1u << DC_SYSTEM);
107 if (lex_match_id ("NEWNAMES"))
110 flip->new_names = parse_variable ();
111 if (!flip->new_names)
115 flip->new_names = dict_lookup_var (default_dict, "CASE_LBL");
121 for (i = 0; i < flip->var_cnt; i++)
122 if (flip->var[i] == flip->new_names)
124 remove_element (flip->var, flip->var_cnt, sizeof *flip->var, i);
130 /* Read the active file into a flip_sink. */
132 temp_trns = temporary = 0;
133 vfm_sink = flip_sink_create (flip);
134 flip->new_names_tail = NULL;
135 procedure (NULL, NULL);
137 /* Flip the data we read. */
140 /* Flip the dictionary. */
141 dict_clear (default_dict);
142 if (!build_dictionary (flip))
144 discard_variables ();
147 flip->case_size = dict_get_case_size (default_dict);
149 /* Set up flipped data for reading. */
150 vfm_source = flip_source_create (flip);
152 return lex_end_of_command ();
155 destroy_flip_pgm (flip);
161 destroy_flip_pgm (struct flip_pgm *flip)
163 struct varname *iter, *next;
166 free (flip->idx_to_fv);
167 for (iter = flip->new_names_head; iter != NULL; iter = next)
172 if (flip->file != NULL)
177 /* Make a new variable with base name NAME, which is bowdlerized and
178 mangled until acceptable, and returns success. */
180 make_new_var (char name[])
182 /* Fix invalid characters. */
186 for (cp = name; *cp && !isspace (*cp); cp++)
188 *cp = toupper ((unsigned char) *cp);
189 if (!isalpha (*cp) && *cp != '@' && *cp != '#'
190 && (cp == name || (*cp != '.' && *cp != '$' && *cp != '_'
194 *cp = 'V'; /* _ not valid in first position. */
202 if (dict_create_var (default_dict, name, 0))
205 /* Add numeric extensions until acceptable. */
207 int len = (int) strlen (name);
211 for (i = 1; i < 10000000; i++)
213 int ofs = min (7 - intlog10 (i), len);
214 memcpy (n, name, ofs);
215 sprintf (&n[ofs], "%d", i);
217 if (dict_create_var (default_dict, n, 0))
222 msg (SE, _("Could not create acceptable variant for variable %s."), name);
226 /* Make a new dictionary for all the new variable names. */
228 build_dictionary (struct flip_pgm *flip)
230 dict_create_var_assert (default_dict, "CASE_LBL", 8);
232 if (flip->new_names_head == NULL)
236 if (flip->case_cnt > 99999)
238 msg (SE, _("Cannot create more than 99999 variable names."));
242 for (i = 0; i < flip->case_cnt; i++)
247 sprintf (s, "VAR%03d", i);
248 v = dict_create_var_assert (default_dict, s, 0);
255 for (v = flip->new_names_head; v; v = v->next)
256 if (!make_new_var (v->name))
263 /* Cases during transposition. */
264 struct flip_sink_info
266 struct flip_pgm *flip; /* FLIP program. */
267 union value *output_buf; /* Case output buffer. */
270 /* Creates a flip sink based on FLIP. */
271 static struct case_sink *
272 flip_sink_create (struct flip_pgm *flip)
274 struct flip_sink_info *info = xmalloc (sizeof *info);
278 info->output_buf = xmalloc (sizeof *info->output_buf * flip->var_cnt);
280 flip->file = tmpfile ();
282 msg (FE, _("Could not create temporary file for FLIP."));
284 /* Write variable names as first case. */
285 for (i = 0; i < flip->var_cnt; i++)
286 st_bare_pad_copy (info->output_buf[i].s, flip->var[i]->name, 8);
287 if (fwrite (info->output_buf, sizeof *info->output_buf,
288 flip->var_cnt, flip->file) != (size_t) flip->var_cnt)
289 msg (FE, _("Error writing FLIP file: %s."), strerror (errno));
293 return create_case_sink (&flip_sink_class, default_dict, info);
296 /* Writes case C to the FLIP sink. */
298 flip_sink_write (struct case_sink *sink, const struct ccase *c)
300 struct flip_sink_info *info = sink->aux;
301 struct flip_pgm *flip = info->flip;
306 if (flip->new_names != NULL)
308 struct varname *v = xmalloc (sizeof (struct varname));
310 if (flip->new_names->type == NUMERIC)
312 double f = case_num (c, flip->idx_to_fv[flip->new_names->index]);
315 strcpy (v->name, "VSYSMIS");
316 else if (f < INT_MIN)
317 strcpy (v->name, "VNEGINF");
318 else if (f > INT_MAX)
319 strcpy (v->name, "VPOSINF");
322 char name[INT_DIGITS + 2];
323 sprintf (name, "V%d", (int) f);
324 strncpy (v->name, name, 8);
330 int width = min (flip->new_names->width, 8);
331 memcpy (v->name, case_str (c, flip->idx_to_fv[flip->new_names->index]),
336 if (flip->new_names_head == NULL)
337 flip->new_names_head = v;
339 flip->new_names_tail->next = v;
340 flip->new_names_tail = v;
343 /* Write to external file. */
344 for (i = 0; i < flip->var_cnt; i++)
348 if (flip->var[i]->type == NUMERIC)
349 out = case_num (c, flip->idx_to_fv[flip->var[i]->index]);
352 info->output_buf[i].f = out;
355 if (fwrite (info->output_buf, sizeof *info->output_buf,
356 flip->var_cnt, flip->file) != (size_t) flip->var_cnt)
357 msg (FE, _("Error writing FLIP file: %s."), strerror (errno));
360 /* Transposes the external file into a new file. */
362 flip_file (struct flip_pgm *flip)
365 size_t case_capacity;
367 union value *input_buf, *output_buf;
368 FILE *input_file, *output_file;
370 /* Allocate memory for many cases. */
371 case_bytes = flip->var_cnt * sizeof *input_buf;
372 case_capacity = get_max_workspace() / case_bytes;
373 if (case_capacity > flip->case_cnt * 2)
374 case_capacity = flip->case_cnt * 2;
375 if (case_capacity < 2)
379 size_t bytes = case_bytes * case_capacity;
380 if (case_capacity > 2)
381 input_buf = malloc (bytes);
383 input_buf = xmalloc (bytes);
384 if (input_buf != NULL)
388 if (case_capacity < 2)
392 /* Use half the allocated memory for input_buf, half for
395 output_buf = input_buf + flip->var_cnt * case_capacity;
397 input_file = flip->file;
398 if (fseek (input_file, 0, SEEK_SET) != 0)
399 msg (FE, _("Error rewinding FLIP file: %s."), strerror (errno));
401 output_file = tmpfile ();
402 if (output_file == NULL)
403 msg (FE, _("Error creating FLIP source file."));
405 for (case_idx = 0; case_idx < flip->case_cnt; )
407 unsigned long read_cases = min (flip->case_cnt - case_idx,
411 if (read_cases != fread (input_buf, case_bytes, read_cases, input_file))
412 msg (FE, _("Error reading FLIP file: %s."), strerror (errno));
414 for (i = 0; i < flip->var_cnt; i++)
418 for (j = 0; j < read_cases; j++)
419 output_buf[j] = input_buf[i + j * flip->var_cnt];
426 #define off_t long int
429 if (fseeko (output_file,
430 sizeof *input_buf * (case_idx
431 + (off_t) i * flip->case_cnt),
433 msg (FE, _("Error seeking FLIP source file: %s."),
436 if (fwrite (output_buf, sizeof *output_buf, read_cases, output_file)
438 msg (FE, _("Error writing FLIP source file: %s."),
442 case_idx += read_cases;
448 if (fseek (output_file, 0, SEEK_SET) != 0)
449 msg (FE, _("Error rewind FLIP source file: %s."), strerror (errno));
450 flip->file = output_file;
453 /* Destroy sink's internal data. */
455 flip_sink_destroy (struct case_sink *sink)
457 struct flip_sink_info *info = sink->aux;
459 free (info->output_buf);
463 /* FLIP sink class. */
464 static const struct case_sink_class flip_sink_class =
473 /* Creates and returns a FLIP source based on PGM,
474 which should have already been used as a sink. */
475 static struct case_source *
476 flip_source_create (struct flip_pgm *pgm)
478 return create_case_source (&flip_source_class, pgm);
481 /* Reads the FLIP stream. Copies each case into C and calls
482 WRITE_CASE passing WC_DATA. */
484 flip_source_read (struct case_source *source,
486 write_case_func *write_case, write_case_data wc_data)
488 struct flip_pgm *flip = source->aux;
489 union value *input_buf;
492 input_buf = xmalloc (sizeof *input_buf * flip->case_cnt);
493 for (i = 0; i < flip->var_cnt; i++)
497 if (fread (input_buf, sizeof *input_buf, flip->case_cnt,
498 flip->file) != flip->case_cnt)
500 if (ferror (flip->file))
501 msg (SE, _("Error reading FLIP temporary file: %s."),
503 else if (feof (flip->file))
504 msg (SE, _("Unexpected end of file reading FLIP temporary file."));
510 for (j = 0; j < flip->case_cnt; j++)
511 case_data_rw (c, j)->f = input_buf[j].f;
512 if (!write_case (wc_data))
518 /* Destroy internal data in SOURCE. */
520 flip_source_destroy (struct case_source *source)
522 struct flip_pgm *flip = source->aux;
524 destroy_flip_pgm (flip);
527 static const struct case_source_class flip_source_class =