1 /* PSPP - computes sample statistics.
2 Copyright (C) 1997-9, 2000 Free Software Foundation, Inc.
4 This program is free software; you can redistribute it and/or
5 modify it under the terms of the GNU General Public License as
6 published by the Free Software Foundation; either version 2 of the
7 License, or (at your option) any later version.
9 This program is distributed in the hope that it will be useful, but
10 WITHOUT ANY WARRANTY; without even the implied warranty of
11 MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
12 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, write to the Free Software
16 Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA
26 #ifdef HAVE_SYS_TYPES_H
27 #include <sys/types.h>
30 #include <data/case-sink.h>
31 #include <data/case-source.h>
32 #include <data/case.h>
33 #include <data/dictionary.h>
34 #include <data/procedure.h>
35 #include <data/settings.h>
36 #include <data/value.h>
37 #include <data/variable.h>
38 #include <language/command.h>
39 #include <language/lexer/lexer.h>
40 #include <language/lexer/variable-parser.h>
41 #include <libpspp/alloc.h>
42 #include <libpspp/array.h>
43 #include <libpspp/assertion.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>
54 #define _(msgid) gettext (msgid)
56 /* List of variable names. */
60 char name[SHORT_NAME_LEN + 1];
63 /* Represents a FLIP input program. */
66 struct pool *pool; /* Pool containing FLIP data. */
67 struct variable **var; /* Variables to transpose. */
68 int *idx_to_fv; /* var[]->index to compacted sink case fv. */
69 size_t var_cnt; /* Number of elements in `var'. */
70 int case_cnt; /* Pre-flip case count. */
71 size_t case_size; /* Post-flip bytes per case. */
73 union value *output_buf; /* Case output buffer. */
75 struct variable *new_names; /* Variable containing new variable names. */
76 struct varname *new_names_head; /* First new variable. */
77 struct varname *new_names_tail; /* Last new variable. */
79 FILE *file; /* Temporary file containing data. */
80 union value *input_buf; /* Input buffer for temporary file. */
81 size_t cases_read; /* Number of cases already read. */
82 bool error; /* Error reading temporary file? */
85 static void destroy_flip_pgm (struct flip_pgm *);
86 static struct case_sink *flip_sink_create (struct dataset *ds, struct flip_pgm *);
87 static struct case_source *flip_source_create (struct flip_pgm *);
88 static bool flip_file (struct flip_pgm *);
89 static int build_dictionary (struct dictionary *, struct flip_pgm *);
91 static const struct case_source_class flip_source_class;
92 static const struct case_sink_class flip_sink_class;
94 /* Parses and executes FLIP. */
96 cmd_flip (struct lexer *lexer, struct dataset *ds)
98 struct flip_pgm *flip;
99 struct case_sink *sink;
100 struct dictionary *dict = dataset_dict (ds);
103 if (proc_make_temporary_transformations_permanent (ds))
104 msg (SW, _("FLIP ignores TEMPORARY. "
105 "Temporary transformations will be made permanent."));
107 flip = pool_create_container (struct flip_pgm, pool);
109 flip->idx_to_fv = dict_get_compacted_dict_index_to_case_index (dict);
110 pool_register (flip->pool, free, flip->idx_to_fv);
113 flip->new_names = NULL;
114 flip->new_names_head = NULL;
115 flip->new_names_tail = NULL;
117 flip->input_buf = NULL;
118 flip->cases_read = 0;
121 lex_match (lexer, '/');
122 if (lex_match_id (lexer, "VARIABLES"))
124 lex_match (lexer, '=');
125 if (!parse_variables (lexer, dict, &flip->var, &flip->var_cnt,
128 lex_match (lexer, '/');
131 dict_get_vars (dict, &flip->var, &flip->var_cnt, 1u << DC_SYSTEM);
132 pool_register (flip->pool, free, flip->var);
134 lex_match (lexer, '/');
135 if (lex_match_id (lexer, "NEWNAMES"))
137 lex_match (lexer, '=');
138 flip->new_names = parse_variable (lexer, dict);
139 if (!flip->new_names)
143 flip->new_names = dict_lookup_var (dict, "CASE_LBL");
149 for (i = 0; i < flip->var_cnt; i++)
150 if (flip->var[i] == flip->new_names)
152 remove_element (flip->var, flip->var_cnt, sizeof *flip->var, i);
158 /* Read the active file into a flip_sink. */
160 proc_make_temporary_transformations_permanent (ds);
161 sink = flip_sink_create (ds, flip);
164 proc_set_sink (ds, sink);
165 flip->new_names_tail = NULL;
166 ok = procedure (ds,NULL, NULL);
168 /* Flip the data we read. */
169 if (!flip_file (flip))
171 discard_variables (ds);
175 /* Flip the dictionary. */
177 if (!build_dictionary (dict, flip))
179 discard_variables (ds);
182 flip->case_size = dict_get_case_size (dict);
184 /* Set up flipped data for reading. */
185 proc_set_source (ds, flip_source_create (flip));
187 return ok ? lex_end_of_command (lexer) : CMD_CASCADING_FAILURE;
190 destroy_flip_pgm (flip);
191 return CMD_CASCADING_FAILURE;
196 destroy_flip_pgm (struct flip_pgm *flip)
199 pool_destroy (flip->pool);
202 /* Make a new variable with base name NAME, which is bowdlerized and
203 mangled until acceptable, and returns success. */
205 make_new_var (struct dictionary *dict, char name[])
209 /* Trim trailing spaces. */
210 cp = strchr (name, '\0');
211 while (cp > name && isspace ((unsigned char) cp[-1]))
214 /* Fix invalid characters. */
215 for (cp = name; *cp && cp < name + SHORT_NAME_LEN; cp++)
218 if (!lex_is_id1 (*cp) || *cp == '$')
223 if (!lex_is_idn (*cp))
227 str_uppercase (name);
229 if (dict_create_var (dict, name, 0))
232 /* Add numeric extensions until acceptable. */
234 const int len = (int) strlen (name);
235 char n[SHORT_NAME_LEN + 1];
238 for (i = 1; i < 10000000; i++)
240 int ofs = MIN (7 - intlog10 (i), len);
241 memcpy (n, name, ofs);
242 sprintf (&n[ofs], "%d", i);
244 if (dict_create_var (dict, n, 0))
249 msg (SE, _("Could not create acceptable variant for variable %s."), name);
253 /* Make a new dictionary for all the new variable names. */
255 build_dictionary (struct dictionary *dict, struct flip_pgm *flip)
257 dict_create_var_assert (dict, "CASE_LBL", 8);
259 if (flip->new_names_head == NULL)
263 if (flip->case_cnt > 99999)
265 msg (SE, _("Cannot create more than 99999 variable names."));
269 for (i = 0; i < flip->case_cnt; i++)
272 char s[SHORT_NAME_LEN + 1];
274 sprintf (s, "VAR%03d", i);
275 v = dict_create_var_assert (dict, s, 0);
282 for (v = flip->new_names_head; v; v = v->next)
283 if (!make_new_var (dict, v->name))
290 /* Creates a flip sink based on FLIP. */
291 static struct case_sink *
292 flip_sink_create (struct dataset *ds, struct flip_pgm *flip)
296 flip->output_buf = pool_nalloc (flip->pool,
297 flip->var_cnt, sizeof *flip->output_buf);
299 flip->file = pool_tmpfile (flip->pool);
300 if (flip->file == NULL)
302 msg (SE, _("Could not create temporary file for FLIP."));
306 /* Write variable names as first case. */
307 for (i = 0; i < flip->var_cnt; i++)
308 buf_copy_str_rpad (flip->output_buf[i].s, MAX_SHORT_STRING,
309 var_get_name (flip->var[i]));
310 if (fwrite (flip->output_buf, sizeof *flip->output_buf,
311 flip->var_cnt, flip->file) != (size_t) flip->var_cnt)
313 msg (SE, _("Error writing FLIP file: %s."), strerror (errno));
319 return create_case_sink (&flip_sink_class,
321 dataset_get_casefile_factory (ds),
325 /* Writes case C to the FLIP sink.
326 Returns true if successful, false if an I/O error occurred. */
328 flip_sink_write (struct case_sink *sink, const struct ccase *c)
330 struct flip_pgm *flip = sink->aux;
335 if (flip->new_names != NULL)
337 struct varname *v = pool_alloc (flip->pool, sizeof *v);
338 int fv = flip->idx_to_fv[var_get_dict_index (flip->new_names)];
340 if (var_is_numeric (flip->new_names))
342 double f = case_num_idx (c, fv);
345 strcpy (v->name, "VSYSMIS");
346 else if (f < INT_MIN)
347 strcpy (v->name, "VNEGINF");
348 else if (f > INT_MAX)
349 strcpy (v->name, "VPOSINF");
351 snprintf (v->name, sizeof v->name, "V%d", (int) f);
355 int width = MIN (var_get_width (flip->new_names), MAX_SHORT_STRING);
356 memcpy (v->name, case_str_idx (c, fv), width);
360 if (flip->new_names_head == NULL)
361 flip->new_names_head = v;
363 flip->new_names_tail->next = v;
364 flip->new_names_tail = v;
367 /* Write to external file. */
368 for (i = 0; i < flip->var_cnt; i++)
372 if (var_is_numeric (flip->var[i]))
374 int fv = flip->idx_to_fv[var_get_dict_index (flip->var[i])];
375 out = case_num_idx (c, fv);
379 flip->output_buf[i].f = out;
382 if (fwrite (flip->output_buf, sizeof *flip->output_buf,
383 flip->var_cnt, flip->file) != (size_t) flip->var_cnt)
385 msg (SE, _("Error writing FLIP file: %s."), strerror (errno));
391 /* Transposes the external file into a new file. */
393 flip_file (struct flip_pgm *flip)
396 size_t case_capacity;
398 union value *input_buf, *output_buf;
399 FILE *input_file, *output_file;
401 /* Allocate memory for many cases. */
402 case_bytes = flip->var_cnt * sizeof *input_buf;
403 case_capacity = get_workspace () / case_bytes;
404 if (case_capacity > flip->case_cnt * 2)
405 case_capacity = flip->case_cnt * 2;
406 if (case_capacity < 2)
410 size_t bytes = case_bytes * case_capacity;
411 if (case_capacity > 2)
412 input_buf = malloc (bytes);
414 input_buf = xmalloc (bytes);
415 if (input_buf != NULL)
419 if (case_capacity < 2)
422 pool_register (flip->pool, free, input_buf);
424 /* Use half the allocated memory for input_buf, half for
427 output_buf = input_buf + flip->var_cnt * case_capacity;
429 input_file = flip->file;
430 if (fseek (input_file, 0, SEEK_SET) != 0)
432 msg (SE, _("Error rewinding FLIP file: %s."), strerror (errno));
436 output_file = pool_tmpfile (flip->pool);
437 if (output_file == NULL)
439 msg (SE, _("Error creating FLIP source file."));
443 for (case_idx = 0; case_idx < flip->case_cnt; )
445 unsigned long read_cases = MIN (flip->case_cnt - case_idx,
449 if (read_cases != fread (input_buf, case_bytes, read_cases, input_file))
451 if (ferror (input_file))
452 msg (SE, _("Error reading FLIP file: %s."), strerror (errno));
454 msg (SE, _("Unexpected end of file reading FLIP file."));
458 for (i = 0; i < flip->var_cnt; i++)
462 for (j = 0; j < read_cases; j++)
463 output_buf[j] = input_buf[i + j * flip->var_cnt];
470 #define off_t long int
473 if (fseeko (output_file,
474 sizeof *input_buf * (case_idx
475 + (off_t) i * flip->case_cnt),
478 msg (SE, _("Error seeking FLIP source file: %s."),
483 if (fwrite (output_buf, sizeof *output_buf, read_cases, output_file)
486 msg (SE, _("Error writing FLIP source file: %s."),
492 case_idx += read_cases;
495 if (pool_fclose (flip->pool, input_file) == EOF)
497 msg (SE, _("Error closing FLIP source file: %s."), strerror (errno));
500 pool_unregister (flip->pool, input_buf);
503 if (fseek (output_file, 0, SEEK_SET) != 0)
505 msg (SE, _("Error rewinding FLIP source file: %s."), strerror (errno));
508 flip->file = output_file;
513 /* FLIP sink class. */
514 static const struct case_sink_class flip_sink_class =
523 /* Creates and returns a FLIP source based on PGM,
524 which should have already been used as a sink. */
525 static struct case_source *
526 flip_source_create (struct flip_pgm *pgm)
528 return create_case_source (&flip_source_class, pgm);
531 /* Reads one case into C.
532 Returns true if successful, false at end of file or if an
533 I/O error occurred. */
535 flip_source_read (struct case_source *source, struct ccase *c)
537 struct flip_pgm *flip = source->aux;
540 if (flip->error || flip->cases_read >= flip->var_cnt)
543 if (flip->input_buf == NULL)
544 flip->input_buf = pool_nmalloc (flip->pool,
545 flip->case_cnt, sizeof *flip->input_buf);
547 if (fread (flip->input_buf, sizeof *flip->input_buf, flip->case_cnt,
548 flip->file) != flip->case_cnt)
550 if (ferror (flip->file))
551 msg (SE, _("Error reading FLIP temporary file: %s."),
553 else if (feof (flip->file))
554 msg (SE, _("Unexpected end of file reading FLIP temporary file."));
561 for (i = 0; i < flip->case_cnt; i++)
562 case_data_rw_idx (c, i)->f = flip->input_buf[i].f;
569 /* Destroys the source.
570 Returns true if successful read, false if an I/O occurred
571 during destruction or previously. */
573 flip_source_destroy (struct case_source *source)
575 struct flip_pgm *flip = source->aux;
576 bool ok = !flip->error;
577 destroy_flip_pgm (flip);
581 static const struct case_source_class flip_source_class =