1 /* PSPP - computes sample statistics.
2 Copyright (C) 1997-9, 2000 Free Software Foundation, Inc.
3 Written by Ben Pfaff <blp@gnu.org>.
5 This program is free software; you can redistribute it and/or
6 modify it under the terms of the GNU General Public License as
7 published by the Free Software Foundation; either version 2 of the
8 License, or (at your option) any later version.
10 This program is distributed in the hope that it will be useful, but
11 WITHOUT ANY WARRANTY; without even the implied warranty of
12 MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
13 General Public License for more details.
15 You should have received a copy of the GNU General Public License
16 along with this program; if not, write to the Free Software
17 Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA
26 #include "dictionary.h"
36 #define _(msgid) gettext (msgid)
38 /* FIXME: Implement PRINT subcommand. */
40 /* Explains how to recode one value. `from' must be first element. */
43 union value from; /* Original value. */
44 double to; /* Recoded value. */
47 /* Explains how to recode an AUTORECODE variable. */
50 struct variable *src; /* Source variable. */
51 struct variable *dest; /* Target variable. */
52 struct hsh_table *items; /* Hash table of `freq's. */
55 /* AUTORECODE transformation. */
56 struct autorecode_trns
59 struct pool *owner; /* Contains AUTORECODE specs. */
60 struct arc_spec *specs; /* AUTORECODE specifications. */
61 size_t spec_cnt; /* Number of specifications. */
64 /* Descending or ascending sort order. */
71 /* AUTORECODE data. */
74 struct variable **src_vars; /* Source variables. */
75 char **dst_names; /* Target variable names. */
76 struct variable **dst_vars; /* Target variables. */
77 struct hsh_table **src_values; /* `union value's of source vars. */
78 size_t var_cnt; /* Number of variables. */
79 struct pool *src_values_pool; /* Pool used by src_values. */
80 enum direction direction; /* Sort order. */
81 int print; /* Print mapping table if nonzero. */
84 static trns_proc_func autorecode_trns_proc;
85 static trns_free_func autorecode_trns_free;
86 static int autorecode_proc_func (struct ccase *, void *);
87 static hsh_compare_func compare_alpha_value, compare_numeric_value;
88 static hsh_hash_func hash_alpha_value, hash_numeric_value;
90 static void recode (const struct autorecode_pgm *);
91 static void arc_free (struct autorecode_pgm *);
93 /* Performs the AUTORECODE procedure. */
97 struct autorecode_pgm arc;
102 arc.dst_names = NULL;
104 arc.src_values = NULL;
106 arc.src_values_pool = NULL;
107 arc.direction = ASCENDING;
111 lex_match_id ("VARIABLES");
113 if (!parse_variables (default_dict, &arc.src_vars, &arc.var_cnt,
116 if (!lex_force_match_id ("INTO"))
119 if (!parse_DATA_LIST_vars (&arc.dst_names, &dst_cnt, PV_NONE))
121 if (dst_cnt != arc.var_cnt)
125 msg (SE, _("Source variable count (%u) does not match "
126 "target variable count (%u)."),
127 (unsigned) arc.var_cnt, (unsigned) dst_cnt);
129 for (i = 0; i < dst_cnt; i++)
130 free (arc.dst_names[i]);
131 free (arc.dst_names);
132 arc.dst_names = NULL;
136 while (lex_match ('/'))
137 if (lex_match_id ("DESCENDING"))
138 arc.direction = DESCENDING;
139 else if (lex_match_id ("PRINT"))
143 lex_error (_("expecting end of command"));
147 for (i = 0; i < arc.var_cnt; i++)
151 if (dict_lookup_var (default_dict, arc.dst_names[i]) != NULL)
153 msg (SE, _("Target variable %s duplicates existing variable %s."),
154 arc.dst_names[i], arc.dst_names[i]);
157 for (j = 0; j < i; j++)
158 if (!strcasecmp (arc.dst_names[i], arc.dst_names[j]))
160 msg (SE, _("Duplicate variable name %s among target variables."),
166 arc.src_values_pool = pool_create ();
167 arc.dst_vars = xnmalloc (arc.var_cnt, sizeof *arc.dst_vars);
168 arc.src_values = xnmalloc (arc.var_cnt, sizeof *arc.src_values);
169 for (i = 0; i < dst_cnt; i++)
170 if (arc.src_vars[i]->type == ALPHA)
171 arc.src_values[i] = hsh_create (10, compare_alpha_value,
172 hash_alpha_value, NULL, arc.src_vars[i]);
174 arc.src_values[i] = hsh_create (10, compare_numeric_value,
175 hash_numeric_value, NULL, NULL);
177 procedure (autorecode_proc_func, &arc);
179 for (i = 0; i < arc.var_cnt; i++)
181 arc.dst_vars[i] = dict_create_var_assert (default_dict,
182 arc.dst_names[i], 0);
183 arc.dst_vars[i]->init = 0;
196 arc_free (struct autorecode_pgm *arc)
198 free (arc->src_vars);
199 if (arc->dst_names != NULL)
203 for (i = 0; i < arc->var_cnt; i++)
204 free (arc->dst_names[i]);
205 free (arc->dst_names);
207 free (arc->dst_vars);
208 if (arc->src_values != NULL)
212 for (i = 0; i < arc->var_cnt; i++)
213 hsh_destroy (arc->src_values[i]);
214 free (arc->src_values);
216 pool_destroy (arc->src_values_pool);
220 /* AUTORECODE transformation. */
223 recode (const struct autorecode_pgm *arc)
225 struct autorecode_trns *t;
229 pool = pool_create ();
230 t = xmalloc (sizeof *t);
231 t->h.proc = autorecode_trns_proc;
232 t->h.free = autorecode_trns_free;
234 t->specs = pool_nalloc (t->owner, arc->var_cnt, sizeof *t->specs);
235 t->spec_cnt = arc->var_cnt;
236 for (i = 0; i < arc->var_cnt; i++)
238 struct arc_spec *spec = &t->specs[i];
239 void *const *p = hsh_sort (arc->src_values[i]);
240 int count = hsh_count (arc->src_values[i]);
243 spec->src = arc->src_vars[i];
244 spec->dest = arc->dst_vars[i];
246 if (arc->src_vars[i]->type == ALPHA)
247 spec->items = hsh_create (2 * count, compare_alpha_value,
248 hash_alpha_value, NULL, arc->src_vars[i]);
250 spec->items = hsh_create (2 * count, compare_numeric_value,
251 hash_numeric_value, NULL, NULL);
253 for (j = 0; *p; p++, j++)
255 struct arc_item *item = pool_alloc (t->owner, sizeof *item);
256 union value *vp = *p;
258 if (arc->src_vars[i]->type == NUMERIC)
259 item->from.f = vp->f;
261 item->from.c = pool_strdup (t->owner, vp->c);
262 item->to = arc->direction == ASCENDING ? j + 1 : count - j;
263 hsh_force_insert (spec->items, item);
266 add_transformation (&t->h);
270 autorecode_trns_proc (struct trns_header * trns, struct ccase * c,
273 struct autorecode_trns *t = (struct autorecode_trns *) trns;
276 for (i = 0; i < t->spec_cnt; i++)
278 struct arc_spec *spec = &t->specs[i];
279 struct arc_item *item;
282 if (spec->src->type == NUMERIC)
283 v.f = case_num (c, spec->src->fv);
285 v.c = (char *) case_str (c, spec->src->fv);
286 item = hsh_force_find (spec->items, &v);
288 case_data_rw (c, spec->dest->fv)->f = item->to;
294 autorecode_trns_free (struct trns_header * trns)
296 struct autorecode_trns *t = (struct autorecode_trns *) trns;
299 for (i = 0; i < t->spec_cnt; i++)
300 hsh_destroy (t->specs[i].items);
301 pool_destroy (t->owner);
304 /* AUTORECODE procedure. */
307 compare_alpha_value (const void *a_, const void *b_, void *v_)
309 const union value *a = a_;
310 const union value *b = b_;
311 const struct variable *v = v_;
313 return memcmp (a->c, b->c, v->width);
317 hash_alpha_value (const void *a_, void *v_)
319 const union value *a = a_;
320 const struct variable *v = v_;
322 return hsh_hash_bytes (a->c, v->width);
326 compare_numeric_value (const void *a_, const void *b_, void *foo UNUSED)
328 const union value *a = a_;
329 const union value *b = b_;
331 return a->f < b->f ? -1 : a->f > b->f;
335 hash_numeric_value (const void *a_, void *foo UNUSED)
337 const union value *a = a_;
339 return hsh_hash_double (a->f);
343 autorecode_proc_func (struct ccase *c, void *arc_)
345 struct autorecode_pgm *arc = arc_;
348 for (i = 0; i < arc->var_cnt; i++)
350 union value v, *vp, **vpp;
352 if (arc->src_vars[i]->type == NUMERIC)
353 v.f = case_num (c, arc->src_vars[i]->fv);
355 v.c = (char *) case_str (c, arc->src_vars[i]->fv);
357 vpp = (union value **) hsh_probe (arc->src_values[i], &v);
360 vp = pool_alloc (arc->src_values_pool, sizeof *vp);
361 if (arc->src_vars[i]->type == NUMERIC)
364 vp->c = pool_strndup (arc->src_values_pool,
365 v.c, arc->src_vars[i]->width);