1 /* PSPP - a program for statistical analysis.
2 Copyright (C) 1997-9, 2000, 2009, 2010 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/>. */
20 #include <data/case.h>
21 #include <data/casereader.h>
22 #include <data/dictionary.h>
23 #include <data/procedure.h>
24 #include <data/transformations.h>
25 #include <data/variable.h>
26 #include <language/command.h>
27 #include <language/lexer/lexer.h>
28 #include <language/lexer/variable-parser.h>
29 #include <libpspp/compiler.h>
30 #include <libpspp/hash.h>
31 #include <libpspp/message.h>
32 #include <libpspp/pool.h>
33 #include <libpspp/str.h>
38 #define _(msgid) gettext (msgid)
40 /* FIXME: Implement PRINT subcommand. */
42 /* An AUTORECODE variable's original value. */
45 double f; /* Numeric. */
46 char *c; /* Short or long string. */
49 /* Explains how to recode one value. `from' must be first element. */
52 union arc_value from; /* Original value. */
53 double to; /* Recoded value. */
56 /* Explains how to recode an AUTORECODE variable. */
59 const struct variable *src; /* Source variable. */
60 struct variable *dest; /* Target variable. */
61 struct hsh_table *items; /* Hash table of `freq's. */
64 /* AUTORECODE transformation. */
65 struct autorecode_trns
67 struct pool *pool; /* Contains AUTORECODE specs. */
68 struct arc_spec *specs; /* AUTORECODE specifications. */
69 size_t spec_cnt; /* Number of specifications. */
72 /* Descending or ascending sort order. */
79 /* AUTORECODE data. */
82 const struct variable **src_vars; /* Source variables. */
83 char **dst_names; /* Target variable names. */
84 struct variable **dst_vars; /* Target variables. */
85 struct hsh_table **src_values; /* `union arc_value's of source vars. */
86 size_t var_cnt; /* Number of variables. */
87 struct pool *src_values_pool; /* Pool used by src_values. */
88 enum direction direction; /* Sort order. */
89 int print; /* Print mapping table if nonzero. */
92 static trns_proc_func autorecode_trns_proc;
93 static trns_free_func autorecode_trns_free;
94 static hsh_compare_func compare_alpha_value, compare_numeric_value;
95 static hsh_hash_func hash_alpha_value, hash_numeric_value;
97 static void recode (struct dataset *, const struct autorecode_pgm *);
98 static void arc_free (struct autorecode_pgm *);
100 /* Performs the AUTORECODE procedure. */
102 cmd_autorecode (struct lexer *lexer, struct dataset *ds)
104 struct autorecode_pgm arc;
105 struct casereader *input;
112 arc.dst_names = NULL;
114 arc.src_values = NULL;
116 arc.src_values_pool = NULL;
117 arc.direction = ASCENDING;
121 lex_match_id (lexer, "VARIABLES");
122 lex_match (lexer, '=');
123 if (!parse_variables_const (lexer, dataset_dict (ds), &arc.src_vars,
124 &arc.var_cnt, PV_NO_DUPLICATE))
126 if (!lex_force_match_id (lexer, "INTO"))
128 lex_match (lexer, '=');
129 if (!parse_DATA_LIST_vars (lexer, &arc.dst_names, &dst_cnt, PV_NONE))
131 if (dst_cnt != arc.var_cnt)
135 msg (SE, _("Source variable count (%zu) does not match "
136 "target variable count (%zu)."),
137 arc.var_cnt, dst_cnt);
139 for (i = 0; i < dst_cnt; i++)
140 free (arc.dst_names[i]);
141 free (arc.dst_names);
142 arc.dst_names = NULL;
146 while (lex_match (lexer, '/'))
147 if (lex_match_id (lexer, "DESCENDING"))
148 arc.direction = DESCENDING;
149 else if (lex_match_id (lexer, "PRINT"))
151 if (lex_token (lexer) != '.')
153 lex_error (lexer, _("expecting end of command"));
157 for (i = 0; i < arc.var_cnt; i++)
161 if (dict_lookup_var (dataset_dict (ds), arc.dst_names[i]) != NULL)
163 msg (SE, _("Target variable %s duplicates existing variable %s."),
164 arc.dst_names[i], arc.dst_names[i]);
167 for (j = 0; j < i; j++)
168 if (!strcasecmp (arc.dst_names[i], arc.dst_names[j]))
170 msg (SE, _("Duplicate variable name %s among target variables."),
176 arc.src_values_pool = pool_create ();
177 arc.dst_vars = xnmalloc (arc.var_cnt, sizeof *arc.dst_vars);
178 arc.src_values = xnmalloc (arc.var_cnt, sizeof *arc.src_values);
179 for (i = 0; i < dst_cnt; i++)
181 /* FIXME: consolodate this hsh_create */
182 if (var_is_alpha (arc.src_vars[i]))
183 arc.src_values[i] = hsh_create (10, compare_alpha_value,
184 hash_alpha_value, NULL, arc.src_vars[i]);
186 arc.src_values[i] = hsh_create (10, compare_numeric_value,
187 hash_numeric_value, NULL, NULL);
190 input = proc_open (ds);
191 for (; (c = casereader_read (input)) != NULL; case_unref (c))
192 for (i = 0; i < arc.var_cnt; i++)
194 union arc_value v, *vp, **vpp;
196 if (var_is_numeric (arc.src_vars[i]))
197 v.f = case_num (c, arc.src_vars[i]);
199 v.c = (char *) case_str (c, arc.src_vars[i]);
201 vpp = (union arc_value **) hsh_probe (arc.src_values[i], &v);
204 vp = pool_alloc (arc.src_values_pool, sizeof *vp);
205 if (var_is_numeric (arc.src_vars[i]))
208 vp->c = pool_clone (arc.src_values_pool,
209 v.c, var_get_width (arc.src_vars[i]));
213 ok = casereader_destroy (input);
214 ok = proc_commit (ds) && ok;
216 for (i = 0; i < arc.var_cnt; i++)
217 arc.dst_vars[i] = dict_create_var_assert (dataset_dict (ds),
218 arc.dst_names[i], 0);
222 return ok ? CMD_SUCCESS : CMD_CASCADING_FAILURE;
226 return CMD_CASCADING_FAILURE;
230 arc_free (struct autorecode_pgm *arc)
232 free (arc->src_vars);
233 if (arc->dst_names != NULL)
237 for (i = 0; i < arc->var_cnt; i++)
238 free (arc->dst_names[i]);
239 free (arc->dst_names);
241 free (arc->dst_vars);
242 if (arc->src_values != NULL)
246 for (i = 0; i < arc->var_cnt; i++)
247 hsh_destroy (arc->src_values[i]);
248 free (arc->src_values);
250 pool_destroy (arc->src_values_pool);
254 /* AUTORECODE transformation. */
257 recode (struct dataset *ds, const struct autorecode_pgm *arc)
259 struct autorecode_trns *trns;
262 trns = pool_create_container (struct autorecode_trns, pool);
263 trns->specs = pool_nalloc (trns->pool, arc->var_cnt, sizeof *trns->specs);
264 trns->spec_cnt = arc->var_cnt;
265 for (i = 0; i < arc->var_cnt; i++)
267 struct arc_spec *spec = &trns->specs[i];
268 void *const *p = hsh_sort (arc->src_values[i]);
269 int count = hsh_count (arc->src_values[i]);
272 spec->src = arc->src_vars[i];
273 spec->dest = arc->dst_vars[i];
275 if (var_is_alpha (arc->src_vars[i]))
276 spec->items = hsh_create (2 * count, compare_alpha_value,
277 hash_alpha_value, NULL, arc->src_vars[i]);
279 spec->items = hsh_create (2 * count, compare_numeric_value,
280 hash_numeric_value, NULL, NULL);
282 for (j = 0; *p; p++, j++)
284 struct arc_item *item = pool_alloc (trns->pool, sizeof *item);
285 union arc_value *vp = *p;
287 if (var_is_numeric (arc->src_vars[i]))
288 item->from.f = vp->f;
290 item->from.c = pool_clone (trns->pool, vp->c,
291 var_get_width (arc->src_vars[i]));
292 item->to = arc->direction == ASCENDING ? j + 1 : count - j;
293 hsh_force_insert (spec->items, item);
296 add_transformation (ds,
297 autorecode_trns_proc, autorecode_trns_free, trns);
300 /* Executes an AUTORECODE transformation. */
302 autorecode_trns_proc (void *trns_, struct ccase **c,
303 casenumber case_idx UNUSED)
305 struct autorecode_trns *trns = trns_;
308 *c = case_unshare (*c);
309 for (i = 0; i < trns->spec_cnt; i++)
311 struct arc_spec *spec = &trns->specs[i];
312 struct arc_item *item;
315 if (var_is_numeric (spec->src))
316 v.f = case_num (*c, spec->src);
318 v.c = (char *) case_str (*c, spec->src);
319 item = hsh_force_find (spec->items, &v);
321 case_data_rw (*c, spec->dest)->f = item->to;
323 return TRNS_CONTINUE;
326 /* Frees an AUTORECODE transformation. */
328 autorecode_trns_free (void *trns_)
330 struct autorecode_trns *trns = trns_;
333 for (i = 0; i < trns->spec_cnt; i++)
334 hsh_destroy (trns->specs[i].items);
335 pool_destroy (trns->pool);
339 /* AUTORECODE procedure. */
342 compare_alpha_value (const void *a_, const void *b_, const void *v_)
344 const union arc_value *a = a_;
345 const union arc_value *b = b_;
346 const struct variable *v = v_;
348 return memcmp (a->c, b->c, var_get_width (v));
352 hash_alpha_value (const void *a_, const void *v_)
354 const union arc_value *a = a_;
355 const struct variable *v = v_;
357 return hash_bytes (a->c, var_get_width (v), 0);
361 compare_numeric_value (const void *a_, const void *b_, const void *aux UNUSED)
363 const union arc_value *a = a_;
364 const union arc_value *b = b_;
366 return a->f < b->f ? -1 : a->f > b->f;
370 hash_numeric_value (const void *a_, const void *aux UNUSED)
372 const union arc_value *a = a_;
374 return hash_double (a->f, 0);