1 /* PSPP - a program for statistical analysis.
2 Copyright (C) 1997-9, 2000, 2009, 2010, 2011, 2013 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/>. */
25 #include "data/case.h"
26 #include "data/casereader.h"
27 #include "data/casereader-provider.h"
28 #include "data/dataset.h"
29 #include "data/dictionary.h"
30 #include "data/settings.h"
31 #include "data/short-names.h"
32 #include "data/value.h"
33 #include "data/variable.h"
34 #include "language/command.h"
35 #include "language/lexer/lexer.h"
36 #include "language/lexer/variable-parser.h"
37 #include "libpspp/array.h"
38 #include "libpspp/assertion.h"
39 #include "libpspp/message.h"
40 #include "libpspp/misc.h"
41 #include "libpspp/pool.h"
42 #include "libpspp/str.h"
43 #include "data/data-in.h"
44 #include "data/data-out.h"
46 #include "gl/intprops.h"
47 #include "gl/minmax.h"
48 #include "gl/xalloc.h"
51 #define _(msgid) gettext (msgid)
53 /* List of variable names. */
57 size_t n_names, allocated_names;
60 static void var_names_init (struct var_names *);
61 static void var_names_add (struct pool *, struct var_names *, const char *);
63 /* Represents a FLIP input program. */
66 struct pool *pool; /* Pool containing FLIP data. */
67 size_t n_vars; /* Pre-flip number of variables. */
68 int n_cases; /* Pre-flip number of cases. */
70 struct variable *new_names_var; /* Variable with new variable names. */
71 const char *encoding; /* Variable names' encoding. */
72 struct var_names old_names; /* Variable names before FLIP. */
73 struct var_names new_names; /* Variable names after FLIP. */
75 FILE *file; /* Temporary file containing data. */
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 void make_new_var (struct dictionary *, const char *name);
86 /* Parses and executes FLIP. */
88 cmd_flip (struct lexer *lexer, struct dataset *ds)
90 struct dictionary *old_dict = dataset_dict (ds);
91 struct dictionary *new_dict = NULL;
92 const struct variable **vars;
93 struct flip_pgm *flip;
94 struct casereader *input, *reader;
99 if (proc_make_temporary_transformations_permanent (ds))
100 lex_ofs_msg (lexer, SW, 0, lex_ofs (lexer) - 1,
102 "Temporary transformations will be made permanent."),
103 "FLIP", "TEMPORARY");
105 flip = pool_create_container (struct flip_pgm, pool);
108 flip->new_names_var = NULL;
109 var_names_init (&flip->old_names);
110 var_names_init (&flip->new_names);
112 flip->cases_read = 0;
115 lex_match (lexer, T_SLASH);
116 if (lex_match_id (lexer, "VARIABLES"))
118 lex_match (lexer, T_EQUALS);
119 if (!parse_variables_const (lexer, old_dict, &vars, &flip->n_vars,
122 lex_match (lexer, T_SLASH);
125 dict_get_vars (old_dict, &vars, &flip->n_vars, DC_SYSTEM);
126 pool_register (flip->pool, free, vars);
128 lex_match (lexer, T_SLASH);
129 if (lex_match_id (lexer, "NEWNAMES"))
131 lex_match (lexer, T_EQUALS);
132 flip->new_names_var = parse_variable (lexer, old_dict);
133 if (!flip->new_names_var)
137 flip->new_names_var = dict_lookup_var (old_dict, "CASE_LBL");
139 if (flip->new_names_var)
141 for (i = 0; i < flip->n_vars; i++)
142 if (vars[i] == flip->new_names_var)
144 remove_element (vars, flip->n_vars, sizeof *vars, i);
149 if (flip->n_vars <= 0)
152 flip->file = pool_create_temp_file (flip->pool);
153 if (flip->file == NULL)
155 msg (SE, _("Could not create temporary file for %s."), "FLIP");
159 /* Save old variable names for use as values of CASE_LBL
160 variable in flipped file. */
161 for (i = 0; i < flip->n_vars; i++)
162 var_names_add (flip->pool, &flip->old_names,
163 pool_strdup (flip->pool, var_get_name (vars[i])));
165 /* Read the active dataset into a flip_sink. */
166 proc_discard_output (ds);
168 /* Save old dictionary. */
169 new_dict = dict_clone (old_dict);
170 flip->encoding = dict_get_encoding (new_dict);
171 dict_clear (new_dict);
173 input = proc_open_filtering (ds, false);
174 while ((c = casereader_read (input)) != NULL)
177 for (i = 0; i < flip->n_vars; i++)
179 const struct variable *v = vars[i];
180 double out = var_is_numeric (v) ? case_num (c, v) : SYSMIS;
181 fwrite (&out, sizeof out, 1, flip->file);
183 if (flip->new_names_var != NULL)
185 const union value *value = case_data (c, flip->new_names_var);
187 if (var_is_numeric (flip->new_names_var))
190 name = (f == SYSMIS ? "VSYSMIS"
191 : f < INT_MIN ? "VNEGINF"
192 : f > INT_MAX ? "VPOSINF"
193 : pool_asprintf (flip->pool, "V%d", (int) f));
197 name = data_out_pool (value, dict_get_encoding (old_dict),
198 var_get_write_format (flip->new_names_var),
199 settings_get_fmt_settings (), flip->pool);
201 var_names_add (flip->pool, &flip->new_names, name);
205 ok = casereader_destroy (input);
206 ok = proc_commit (ds) && ok;
208 /* Flip the data we read. */
209 if (!ok || !flip_file (flip))
215 /* Flip the dictionary. */
216 dict_create_var_assert (new_dict, "CASE_LBL", 8);
217 for (i = 0; i < flip->n_cases; i++)
218 if (flip->new_names.n_names)
219 make_new_var (new_dict, flip->new_names.names[i]);
222 char s[3 + INT_STRLEN_BOUND (i) + 1];
223 sprintf (s, "VAR%03zu", i);
224 dict_create_var_assert (new_dict, s, 0);
227 /* Set up flipped data for reading. */
228 reader = casereader_create_sequential (NULL, dict_get_proto (new_dict),
230 &flip_casereader_class, flip);
231 dataset_set_dict (ds, new_dict);
232 dataset_set_source (ds, reader);
236 dict_unref (new_dict);
237 destroy_flip_pgm (flip);
238 return CMD_CASCADING_FAILURE;
243 destroy_flip_pgm (struct flip_pgm *flip)
246 pool_destroy (flip->pool);
249 /* Make a new variable with base name NAME, which is bowdlerized and
250 mangled until acceptable. */
252 make_new_var (struct dictionary *dict, const char *name_)
254 char *name = xstrdup (name_);
257 /* Trim trailing spaces. */
258 cp = strchr (name, '\0');
259 while (cp > name && isspace ((unsigned char) cp[-1]))
262 /* Fix invalid characters. */
263 for (cp = name; *cp && cp < name + ID_MAX_LEN; cp++)
266 if (!lex_is_id1 (*cp) || *cp == '$')
271 if (!lex_is_idn (*cp))
276 if (strlen (name) == 0)
279 name = xstrdup ("v");
282 /* Use the mangled name, if it is available, or add numeric
283 extensions until we find one that is. */
284 if (!id_is_plausible (name) || !dict_create_var (dict, name, 0))
286 int len = strlen (name);
290 char n[ID_MAX_LEN + 1];
291 int ofs = MIN (ID_MAX_LEN - 1 - intlog10 (i), len);
292 strncpy (n, name, ofs);
293 sprintf (&n[ofs], "%d", i);
295 if (id_is_plausible (n) && dict_create_var (dict, n, 0))
302 /* Transposes the external file into a new file. */
304 flip_file (struct flip_pgm *flip)
307 size_t case_capacity;
309 double *input_buf, *output_buf;
310 FILE *input_file, *output_file;
312 /* Allocate memory for many cases. */
313 case_bytes = flip->n_vars * sizeof *input_buf;
314 case_capacity = settings_get_workspace () / case_bytes;
315 if (case_capacity > flip->n_cases * 2)
316 case_capacity = flip->n_cases * 2;
317 if (case_capacity < 2)
321 size_t bytes = case_bytes * case_capacity;
322 if (case_capacity > 2)
323 input_buf = malloc (bytes);
325 input_buf = xmalloc (bytes);
326 if (input_buf != NULL)
330 if (case_capacity < 2)
333 pool_register (flip->pool, free, input_buf);
335 /* Use half the allocated memory for input_buf, half for
338 output_buf = input_buf + flip->n_vars * case_capacity;
340 input_file = flip->file;
341 if (fseeko (input_file, 0, SEEK_SET) != 0)
343 msg (SE, _("Error rewinding %s file: %s."), "FLIP", strerror (errno));
347 output_file = pool_create_temp_file (flip->pool);
348 if (output_file == NULL)
350 msg (SE, _("Error creating %s source file."), "FLIP");
354 for (case_idx = 0; case_idx < flip->n_cases;)
356 unsigned long read_cases = MIN (flip->n_cases - case_idx,
360 if (read_cases != fread (input_buf, case_bytes, read_cases, input_file))
362 if (ferror (input_file))
363 msg (SE, _("Error reading %s file: %s."), "FLIP", strerror (errno));
365 msg (SE, _("Unexpected end of file reading %s file."), "FLIP");
369 for (i = 0; i < flip->n_vars; i++)
373 for (j = 0; j < read_cases; j++)
374 output_buf[j] = input_buf[i + j * flip->n_vars];
376 if (fseeko (output_file,
377 sizeof *input_buf * (case_idx
378 + (off_t) i * flip->n_cases),
381 msg (SE, _("Error seeking %s source file: %s."), "FLIP",
386 if (fwrite (output_buf, sizeof *output_buf, read_cases, output_file)
389 msg (SE, _("Error writing %s source file: %s."), "FLIP",
395 case_idx += read_cases;
398 pool_fclose_temp_file (flip->pool, input_file);
399 pool_unregister (flip->pool, input_buf);
402 if (fseeko (output_file, 0, SEEK_SET) != 0)
404 msg (SE, _("Error rewinding %s source file: %s."), "FLIP", strerror (errno));
407 flip->file = output_file;
412 /* Reads and returns one case.
413 Returns a null pointer at end of file or if an I/O error occurred. */
414 static struct ccase *
415 flip_casereader_read (struct casereader *reader, void *flip_)
417 struct flip_pgm *flip = flip_;
421 if (flip->error || flip->cases_read >= flip->n_vars)
424 c = case_create (casereader_get_proto (reader));
425 data_in (ss_cstr (flip->old_names.names[flip->cases_read]), flip->encoding,
426 FMT_A, settings_get_fmt_settings (), case_data_rw_idx (c, 0),
429 for (i = 0; i < flip->n_cases; i++)
432 if (fread (&in, sizeof in, 1, flip->file) != 1)
435 if (ferror (flip->file))
436 msg (SE, _("Error reading %s temporary file: %s."), "FLIP",
438 else if (feof (flip->file))
439 msg (SE, _("Unexpected end of file reading %s temporary file."), "FLIP");
445 *case_num_rw_idx (c, i + 1) = in;
453 /* Destroys the source.
454 Returns true if successful read, false if an I/O occurred
455 during destruction or previously. */
457 flip_casereader_destroy (struct casereader *reader, void *flip_)
459 struct flip_pgm *flip = flip_;
461 casereader_force_error (reader);
462 destroy_flip_pgm (flip);
465 static const struct casereader_class flip_casereader_class =
467 flip_casereader_read,
468 flip_casereader_destroy,
474 var_names_init (struct var_names *vn)
478 vn->allocated_names = 0;
482 var_names_add (struct pool *pool, struct var_names *vn, const char *name)
484 if (vn->n_names >= vn->allocated_names)
485 vn->names = pool_2nrealloc (pool, vn->names, &vn->allocated_names,
487 vn->names[vn->n_names++] = name;