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
22 #include <data/case.h>
23 #include <data/dictionary.h>
24 #include <data/procedure.h>
25 #include <data/transformations.h>
26 #include <data/variable.h>
27 #include <language/command.h>
28 #include <language/lexer/lexer.h>
29 #include <language/lexer/variable-parser.h>
30 #include <libpspp/alloc.h>
31 #include <libpspp/compiler.h>
32 #include <libpspp/hash.h>
33 #include <libpspp/message.h>
34 #include <libpspp/message.h>
35 #include <libpspp/pool.h>
36 #include <libpspp/str.h>
39 #define _(msgid) gettext (msgid)
41 /* FIXME: Implement PRINT subcommand. */
43 /* An AUTORECODE variable's original value. */
46 double f; /* Numeric. */
47 char *c; /* Short or long string. */
50 /* Explains how to recode one value. `from' must be first element. */
53 union arc_value from; /* Original value. */
54 double to; /* Recoded value. */
57 /* Explains how to recode an AUTORECODE variable. */
60 struct variable *src; /* Source variable. */
61 struct variable *dest; /* Target variable. */
62 struct hsh_table *items; /* Hash table of `freq's. */
65 /* AUTORECODE transformation. */
66 struct autorecode_trns
68 struct pool *pool; /* Contains AUTORECODE specs. */
69 struct arc_spec *specs; /* AUTORECODE specifications. */
70 size_t spec_cnt; /* Number of specifications. */
73 /* Descending or ascending sort order. */
80 /* AUTORECODE data. */
83 struct variable **src_vars; /* Source variables. */
84 char **dst_names; /* Target variable names. */
85 struct variable **dst_vars; /* Target variables. */
86 struct hsh_table **src_values; /* `union arc_value's of source vars. */
87 size_t var_cnt; /* Number of variables. */
88 struct pool *src_values_pool; /* Pool used by src_values. */
89 enum direction direction; /* Sort order. */
90 int print; /* Print mapping table if nonzero. */
93 static trns_proc_func autorecode_trns_proc;
94 static trns_free_func autorecode_trns_free;
95 static bool autorecode_proc_func (const struct ccase *, void *, const struct dataset *);
96 static hsh_compare_func compare_alpha_value, compare_numeric_value;
97 static hsh_hash_func hash_alpha_value, hash_numeric_value;
99 static void recode (struct dataset *, const struct autorecode_pgm *);
100 static void arc_free (struct autorecode_pgm *);
102 /* Performs the AUTORECODE procedure. */
104 cmd_autorecode (struct lexer *lexer, struct dataset *ds)
106 struct autorecode_pgm arc;
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 (lexer, dataset_dict (ds), &arc.src_vars, &arc.var_cnt,
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 (%u) does not match "
136 "target variable count (%u)."),
137 (unsigned) arc.var_cnt, (unsigned) 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++)
180 if (var_is_alpha (arc.src_vars[i]))
181 arc.src_values[i] = hsh_create (10, compare_alpha_value,
182 hash_alpha_value, NULL, arc.src_vars[i]);
184 arc.src_values[i] = hsh_create (10, compare_numeric_value,
185 hash_numeric_value, NULL, NULL);
187 ok = procedure (ds, autorecode_proc_func, &arc);
189 for (i = 0; i < arc.var_cnt; i++)
190 arc.dst_vars[i] = dict_create_var_assert (dataset_dict (ds),
191 arc.dst_names[i], 0);
195 return ok ? CMD_SUCCESS : CMD_CASCADING_FAILURE;
199 return CMD_CASCADING_FAILURE;
203 arc_free (struct autorecode_pgm *arc)
205 free (arc->src_vars);
206 if (arc->dst_names != NULL)
210 for (i = 0; i < arc->var_cnt; i++)
211 free (arc->dst_names[i]);
212 free (arc->dst_names);
214 free (arc->dst_vars);
215 if (arc->src_values != NULL)
219 for (i = 0; i < arc->var_cnt; i++)
220 hsh_destroy (arc->src_values[i]);
221 free (arc->src_values);
223 pool_destroy (arc->src_values_pool);
227 /* AUTORECODE transformation. */
230 recode (struct dataset *ds, const struct autorecode_pgm *arc)
232 struct autorecode_trns *trns;
235 trns = pool_create_container (struct autorecode_trns, pool);
236 trns->specs = pool_nalloc (trns->pool, arc->var_cnt, sizeof *trns->specs);
237 trns->spec_cnt = arc->var_cnt;
238 for (i = 0; i < arc->var_cnt; i++)
240 struct arc_spec *spec = &trns->specs[i];
241 void *const *p = hsh_sort (arc->src_values[i]);
242 int count = hsh_count (arc->src_values[i]);
245 spec->src = arc->src_vars[i];
246 spec->dest = arc->dst_vars[i];
248 if (var_is_alpha (arc->src_vars[i]))
249 spec->items = hsh_create (2 * count, compare_alpha_value,
250 hash_alpha_value, NULL, arc->src_vars[i]);
252 spec->items = hsh_create (2 * count, compare_numeric_value,
253 hash_numeric_value, NULL, NULL);
255 for (j = 0; *p; p++, j++)
257 struct arc_item *item = pool_alloc (trns->pool, sizeof *item);
258 union arc_value *vp = *p;
260 if (var_is_numeric (arc->src_vars[i]))
261 item->from.f = vp->f;
263 item->from.c = pool_clone (trns->pool, vp->c,
264 var_get_width (arc->src_vars[i]));
265 item->to = arc->direction == ASCENDING ? j + 1 : count - j;
266 hsh_force_insert (spec->items, item);
269 add_transformation (ds,
270 autorecode_trns_proc, autorecode_trns_free, trns);
273 /* Executes an AUTORECODE transformation. */
275 autorecode_trns_proc (void *trns_, struct ccase *c, casenumber case_idx UNUSED)
277 struct autorecode_trns *trns = trns_;
280 for (i = 0; i < trns->spec_cnt; i++)
282 struct arc_spec *spec = &trns->specs[i];
283 struct arc_item *item;
286 if (var_is_numeric (spec->src))
287 v.f = case_num (c, spec->src);
289 v.c = (char *) case_str (c, spec->src);
290 item = hsh_force_find (spec->items, &v);
292 case_data_rw (c, spec->dest)->f = item->to;
294 return TRNS_CONTINUE;
297 /* Frees an AUTORECODE transformation. */
299 autorecode_trns_free (void *trns_)
301 struct autorecode_trns *trns = trns_;
304 for (i = 0; i < trns->spec_cnt; i++)
305 hsh_destroy (trns->specs[i].items);
306 pool_destroy (trns->pool);
310 /* AUTORECODE procedure. */
313 compare_alpha_value (const void *a_, const void *b_, const void *v_)
315 const union arc_value *a = a_;
316 const union arc_value *b = b_;
317 const struct variable *v = v_;
319 return memcmp (a->c, b->c, var_get_width (v));
323 hash_alpha_value (const void *a_, const void *v_)
325 const union arc_value *a = a_;
326 const struct variable *v = v_;
328 return hsh_hash_bytes (a->c, var_get_width (v));
332 compare_numeric_value (const void *a_, const void *b_, const void *aux UNUSED)
334 const union arc_value *a = a_;
335 const union arc_value *b = b_;
337 return a->f < b->f ? -1 : a->f > b->f;
341 hash_numeric_value (const void *a_, const void *aux UNUSED)
343 const union arc_value *a = a_;
345 return hsh_hash_double (a->f);
349 autorecode_proc_func (const struct ccase *c, void *arc_, const struct dataset *ds UNUSED)
351 struct autorecode_pgm *arc = arc_;
354 for (i = 0; i < arc->var_cnt; i++)
356 union arc_value v, *vp, **vpp;
358 if (var_is_numeric (arc->src_vars[i]))
359 v.f = case_num (c, arc->src_vars[i]);
361 v.c = (char *) case_str (c, arc->src_vars[i]);
363 vpp = (union arc_value **) hsh_probe (arc->src_values[i], &v);
366 vp = pool_alloc (arc->src_values_pool, sizeof *vp);
367 if (var_is_numeric (arc->src_vars[i]))
370 vp->c = pool_clone (arc->src_values_pool,
371 v.c, var_get_width (arc->src_vars[i]));