1 /* PSPP - a program for statistical analysis.
2 Copyright (C) 1997-9, 2000 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/>. */
24 #ifdef HAVE_SYS_TYPES_H
25 #include <sys/types.h>
28 #include <data/case.h>
29 #include <data/casereader.h>
30 #include <data/casereader-provider.h>
31 #include <data/dictionary.h>
32 #include <data/procedure.h>
33 #include <data/settings.h>
34 #include <data/value.h>
35 #include <data/variable.h>
36 #include <language/command.h>
37 #include <language/lexer/lexer.h>
38 #include <language/lexer/variable-parser.h>
39 #include <libpspp/alloc.h>
40 #include <libpspp/array.h>
41 #include <libpspp/assertion.h>
42 #include <libpspp/message.h>
43 #include <libpspp/misc.h>
44 #include <libpspp/pool.h>
45 #include <libpspp/str.h>
51 #define _(msgid) gettext (msgid)
53 /* List of variable names. */
57 char name[SHORT_NAME_LEN + 1];
60 /* Represents a FLIP input program. */
63 struct pool *pool; /* Pool containing FLIP data. */
64 const struct variable **var; /* Variables to transpose. */
65 int *idx_to_fv; /* var[]->index to compacted sink case fv. */
66 size_t var_cnt; /* Number of elements in `var'. */
67 int case_cnt; /* Pre-flip case count. */
68 size_t case_size; /* Post-flip bytes per case. */
70 struct variable *new_names; /* Variable containing new variable names. */
71 struct varname *new_names_head; /* First new variable. */
72 struct varname *new_names_tail; /* Last new variable. */
74 FILE *file; /* Temporary file containing data. */
75 union value *input_buf; /* Input buffer for temporary file. */
76 size_t cases_read; /* Number of cases already read. */
77 bool error; /* Error reading temporary file? */
80 static const struct casereader_class flip_casereader_class;
82 static void destroy_flip_pgm (struct flip_pgm *);
83 static bool flip_file (struct flip_pgm *);
84 static bool build_dictionary (struct dictionary *, struct flip_pgm *);
85 static bool write_flip_case (struct flip_pgm *, const struct ccase *);
87 /* Parses and executes FLIP. */
89 cmd_flip (struct lexer *lexer, struct dataset *ds)
91 struct dictionary *dict = dataset_dict (ds);
92 struct flip_pgm *flip;
93 struct casereader *input, *reader;
94 union value *output_buf;
99 if (proc_make_temporary_transformations_permanent (ds))
100 msg (SW, _("FLIP ignores TEMPORARY. "
101 "Temporary transformations will be made permanent."));
103 flip = pool_create_container (struct flip_pgm, pool);
105 flip->idx_to_fv = dict_get_compacted_dict_index_to_case_index (dict);
106 pool_register (flip->pool, free, flip->idx_to_fv);
109 flip->new_names = NULL;
110 flip->new_names_head = NULL;
111 flip->new_names_tail = NULL;
113 flip->input_buf = NULL;
114 flip->cases_read = 0;
117 lex_match (lexer, '/');
118 if (lex_match_id (lexer, "VARIABLES"))
120 lex_match (lexer, '=');
121 if (!parse_variables_const (lexer, dict, &flip->var, &flip->var_cnt,
124 lex_match (lexer, '/');
127 dict_get_vars (dict, &flip->var, &flip->var_cnt, 1u << DC_SYSTEM);
128 pool_register (flip->pool, free, flip->var);
130 lex_match (lexer, '/');
131 if (lex_match_id (lexer, "NEWNAMES"))
133 lex_match (lexer, '=');
134 flip->new_names = parse_variable (lexer, dict);
135 if (!flip->new_names)
139 flip->new_names = dict_lookup_var (dict, "CASE_LBL");
143 for (i = 0; i < flip->var_cnt; i++)
144 if (flip->var[i] == flip->new_names)
146 remove_element (flip->var, flip->var_cnt, sizeof *flip->var, i);
152 output_buf = pool_nalloc (flip->pool,
153 flip->var_cnt, sizeof *output_buf);
155 flip->file = pool_tmpfile (flip->pool);
156 if (flip->file == NULL)
158 msg (SE, _("Could not create temporary file for FLIP."));
162 /* Write variable names as first case. */
163 for (i = 0; i < flip->var_cnt; i++)
164 buf_copy_str_rpad (output_buf[i].s, MAX_SHORT_STRING,
165 var_get_name (flip->var[i]));
166 if (fwrite (output_buf, sizeof *output_buf,
167 flip->var_cnt, flip->file) != (size_t) flip->var_cnt)
169 msg (SE, _("Error writing FLIP file: %s."), strerror (errno));
175 /* Read the active file into a flip_sink. */
176 proc_make_temporary_transformations_permanent (ds);
177 proc_discard_output (ds);
179 input = proc_open (ds);
180 while (casereader_read (input, &c))
182 write_flip_case (flip, &c);
185 ok = casereader_destroy (input);
186 ok = proc_commit (ds) && ok;
188 /* Flip the data we read. */
189 if (!ok || !flip_file (flip))
191 proc_discard_active_file (ds);
195 /* Flip the dictionary. */
197 if (!build_dictionary (dict, flip))
199 proc_discard_active_file (ds);
202 flip->case_size = dict_get_case_size (dict);
204 /* Set up flipped data for reading. */
205 reader = casereader_create_sequential (NULL, dict_get_next_value_idx (dict),
207 &flip_casereader_class, flip);
208 proc_set_active_file_data (ds, reader);
209 return lex_end_of_command (lexer);
212 destroy_flip_pgm (flip);
213 return CMD_CASCADING_FAILURE;
218 destroy_flip_pgm (struct flip_pgm *flip)
221 pool_destroy (flip->pool);
224 /* Make a new variable with base name NAME, which is bowdlerized and
225 mangled until acceptable, and returns success. */
227 make_new_var (struct dictionary *dict, char name[])
231 /* Trim trailing spaces. */
232 cp = strchr (name, '\0');
233 while (cp > name && isspace ((unsigned char) cp[-1]))
236 /* Fix invalid characters. */
237 for (cp = name; *cp && cp < name + SHORT_NAME_LEN; cp++)
240 if (!lex_is_id1 (*cp) || *cp == '$')
245 if (!lex_is_idn (*cp))
249 str_uppercase (name);
251 if (dict_create_var (dict, name, 0))
254 /* Add numeric extensions until acceptable. */
256 const int len = (int) strlen (name);
257 char n[SHORT_NAME_LEN + 1];
260 for (i = 1; i < 10000000; i++)
262 int ofs = MIN (7 - intlog10 (i), len);
263 memcpy (n, name, ofs);
264 sprintf (&n[ofs], "%d", i);
266 if (dict_create_var (dict, n, 0))
271 msg (SE, _("Could not create acceptable variant for variable %s."), name);
275 /* Make a new dictionary for all the new variable names. */
277 build_dictionary (struct dictionary *dict, struct flip_pgm *flip)
279 dict_create_var_assert (dict, "CASE_LBL", 8);
281 if (flip->new_names_head == NULL)
285 if (flip->case_cnt > 99999)
287 msg (SE, _("Cannot create more than 99999 variable names."));
291 for (i = 0; i < flip->case_cnt; i++)
294 char s[SHORT_NAME_LEN + 1];
296 sprintf (s, "VAR%03d", i);
297 v = dict_create_var_assert (dict, s, 0);
304 for (v = flip->new_names_head; v; v = v->next)
305 if (!make_new_var (dict, v->name))
312 /* Writes case C to the FLIP sink.
313 Returns true if successful, false if an I/O error occurred. */
315 write_flip_case (struct flip_pgm *flip, const struct ccase *c)
321 if (flip->new_names != NULL)
323 struct varname *v = pool_alloc (flip->pool, sizeof *v);
324 int fv = flip->idx_to_fv[var_get_dict_index (flip->new_names)];
326 if (var_is_numeric (flip->new_names))
328 double f = case_num_idx (c, fv);
331 strcpy (v->name, "VSYSMIS");
332 else if (f < INT_MIN)
333 strcpy (v->name, "VNEGINF");
334 else if (f > INT_MAX)
335 strcpy (v->name, "VPOSINF");
337 snprintf (v->name, sizeof v->name, "V%d", (int) f);
341 int width = MIN (var_get_width (flip->new_names), MAX_SHORT_STRING);
342 memcpy (v->name, case_str_idx (c, fv), width);
346 if (flip->new_names_head == NULL)
347 flip->new_names_head = v;
349 flip->new_names_tail->next = v;
350 flip->new_names_tail = v;
353 /* Write to external file. */
354 for (i = 0; i < flip->var_cnt; i++)
358 if (var_is_numeric (flip->var[i]))
360 int fv = flip->idx_to_fv[var_get_dict_index (flip->var[i])];
361 out = case_num_idx (c, fv);
365 fwrite (&out, sizeof out, 1, flip->file);
370 /* Transposes the external file into a new file. */
372 flip_file (struct flip_pgm *flip)
375 size_t case_capacity;
377 union value *input_buf, *output_buf;
378 FILE *input_file, *output_file;
380 /* Allocate memory for many cases. */
381 case_bytes = flip->var_cnt * sizeof *input_buf;
382 case_capacity = get_workspace () / case_bytes;
383 if (case_capacity > flip->case_cnt * 2)
384 case_capacity = flip->case_cnt * 2;
385 if (case_capacity < 2)
389 size_t bytes = case_bytes * case_capacity;
390 if (case_capacity > 2)
391 input_buf = malloc (bytes);
393 input_buf = xmalloc (bytes);
394 if (input_buf != NULL)
398 if (case_capacity < 2)
401 pool_register (flip->pool, free, input_buf);
403 /* Use half the allocated memory for input_buf, half for
406 output_buf = input_buf + flip->var_cnt * case_capacity;
408 input_file = flip->file;
409 if (fseek (input_file, 0, SEEK_SET) != 0)
411 msg (SE, _("Error rewinding FLIP file: %s."), strerror (errno));
415 output_file = pool_tmpfile (flip->pool);
416 if (output_file == NULL)
418 msg (SE, _("Error creating FLIP source file."));
422 for (case_idx = 0; case_idx < flip->case_cnt; )
424 unsigned long read_cases = MIN (flip->case_cnt - case_idx,
428 if (read_cases != fread (input_buf, case_bytes, read_cases, input_file))
430 if (ferror (input_file))
431 msg (SE, _("Error reading FLIP file: %s."), strerror (errno));
433 msg (SE, _("Unexpected end of file reading FLIP file."));
437 for (i = 0; i < flip->var_cnt; i++)
441 for (j = 0; j < read_cases; j++)
442 output_buf[j] = input_buf[i + j * flip->var_cnt];
449 #define off_t long int
452 if (fseeko (output_file,
453 sizeof *input_buf * (case_idx
454 + (off_t) i * flip->case_cnt),
457 msg (SE, _("Error seeking FLIP source file: %s."),
462 if (fwrite (output_buf, sizeof *output_buf, read_cases, output_file)
465 msg (SE, _("Error writing FLIP source file: %s."),
471 case_idx += read_cases;
474 if (pool_fclose (flip->pool, input_file) == EOF)
476 msg (SE, _("Error closing FLIP source file: %s."), strerror (errno));
479 pool_unregister (flip->pool, input_buf);
482 if (fseek (output_file, 0, SEEK_SET) != 0)
484 msg (SE, _("Error rewinding FLIP source file: %s."), strerror (errno));
487 flip->file = output_file;
492 /* Reads one case into C.
493 Returns true if successful, false at end of file or if an
494 I/O error occurred. */
496 flip_casereader_read (struct casereader *reader UNUSED, void *flip_,
499 struct flip_pgm *flip = flip_;
502 if (flip->error || flip->cases_read >= flip->var_cnt)
505 case_create (c, flip->case_cnt);
506 for (i = 0; i < flip->case_cnt; i++)
509 if (fread (&in, sizeof in, 1, flip->file) != 1)
512 if (ferror (flip->file))
513 msg (SE, _("Error reading FLIP temporary file: %s."),
515 else if (feof (flip->file))
516 msg (SE, _("Unexpected end of file reading FLIP temporary file."));
522 case_data_rw_idx (c, i)->f = in;
530 /* Destroys the source.
531 Returns true if successful read, false if an I/O occurred
532 during destruction or previously. */
534 flip_casereader_destroy (struct casereader *reader UNUSED, void *flip_)
536 struct flip_pgm *flip = flip_;
538 casereader_force_error (reader);
539 destroy_flip_pgm (flip);
542 static const struct casereader_class flip_casereader_class =
544 flip_casereader_read,
545 flip_casereader_destroy,