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
23 #include <data/case.h>
24 #include <data/dictionary.h>
25 #include <data/procedure.h>
26 #include <data/transformations.h>
27 #include <data/variable.h>
28 #include <language/command.h>
29 #include <language/lexer/lexer.h>
30 #include <language/lexer/variable-parser.h>
31 #include <libpspp/alloc.h>
32 #include <libpspp/compiler.h>
33 #include <libpspp/hash.h>
34 #include <libpspp/message.h>
35 #include <libpspp/message.h>
36 #include <libpspp/pool.h>
37 #include <libpspp/str.h>
40 #define _(msgid) gettext (msgid)
42 /* FIXME: Implement PRINT subcommand. */
44 /* An AUTORECODE variable's original value. */
47 double f; /* Numeric. */
48 char *c; /* Short or long string. */
51 /* Explains how to recode one value. `from' must be first element. */
54 union arc_value from; /* Original value. */
55 double to; /* Recoded value. */
58 /* Explains how to recode an AUTORECODE variable. */
61 struct variable *src; /* Source variable. */
62 struct variable *dest; /* Target variable. */
63 struct hsh_table *items; /* Hash table of `freq's. */
66 /* AUTORECODE transformation. */
67 struct autorecode_trns
69 struct pool *pool; /* Contains AUTORECODE specs. */
70 struct arc_spec *specs; /* AUTORECODE specifications. */
71 size_t spec_cnt; /* Number of specifications. */
74 /* Descending or ascending sort order. */
81 /* AUTORECODE data. */
84 struct variable **src_vars; /* Source variables. */
85 char **dst_names; /* Target variable names. */
86 struct variable **dst_vars; /* Target variables. */
87 struct hsh_table **src_values; /* `union arc_value's of source vars. */
88 size_t var_cnt; /* Number of variables. */
89 struct pool *src_values_pool; /* Pool used by src_values. */
90 enum direction direction; /* Sort order. */
91 int print; /* Print mapping table if nonzero. */
94 static trns_proc_func autorecode_trns_proc;
95 static trns_free_func autorecode_trns_free;
96 static bool autorecode_proc_func (const struct ccase *, void *, const struct dataset *);
97 static hsh_compare_func compare_alpha_value, compare_numeric_value;
98 static hsh_hash_func hash_alpha_value, hash_numeric_value;
100 static void recode (struct dataset *, const struct autorecode_pgm *);
101 static void arc_free (struct autorecode_pgm *);
103 /* Performs the AUTORECODE procedure. */
105 cmd_autorecode (struct lexer *lexer, struct dataset *ds)
107 struct autorecode_pgm arc;
113 arc.dst_names = NULL;
115 arc.src_values = NULL;
117 arc.src_values_pool = NULL;
118 arc.direction = ASCENDING;
122 lex_match_id (lexer, "VARIABLES");
123 lex_match (lexer, '=');
124 if (!parse_variables (lexer, dataset_dict (ds), &arc.src_vars, &arc.var_cnt,
127 if (!lex_force_match_id (lexer, "INTO"))
129 lex_match (lexer, '=');
130 if (!parse_DATA_LIST_vars (lexer, &arc.dst_names, &dst_cnt, PV_NONE))
132 if (dst_cnt != arc.var_cnt)
136 msg (SE, _("Source variable count (%u) does not match "
137 "target variable count (%u)."),
138 (unsigned) arc.var_cnt, (unsigned) dst_cnt);
140 for (i = 0; i < dst_cnt; i++)
141 free (arc.dst_names[i]);
142 free (arc.dst_names);
143 arc.dst_names = NULL;
147 while (lex_match (lexer, '/'))
148 if (lex_match_id (lexer, "DESCENDING"))
149 arc.direction = DESCENDING;
150 else if (lex_match_id (lexer, "PRINT"))
152 if (lex_token (lexer) != '.')
154 lex_error (lexer, _("expecting end of command"));
158 for (i = 0; i < arc.var_cnt; i++)
162 if (dict_lookup_var (dataset_dict (ds), arc.dst_names[i]) != NULL)
164 msg (SE, _("Target variable %s duplicates existing variable %s."),
165 arc.dst_names[i], arc.dst_names[i]);
168 for (j = 0; j < i; j++)
169 if (!strcasecmp (arc.dst_names[i], arc.dst_names[j]))
171 msg (SE, _("Duplicate variable name %s among target variables."),
177 arc.src_values_pool = pool_create ();
178 arc.dst_vars = xnmalloc (arc.var_cnt, sizeof *arc.dst_vars);
179 arc.src_values = xnmalloc (arc.var_cnt, sizeof *arc.src_values);
180 for (i = 0; i < dst_cnt; i++)
181 if (var_is_alpha (arc.src_vars[i]))
182 arc.src_values[i] = hsh_create (10, compare_alpha_value,
183 hash_alpha_value, NULL, arc.src_vars[i]);
185 arc.src_values[i] = hsh_create (10, compare_numeric_value,
186 hash_numeric_value, NULL, NULL);
188 ok = procedure (ds, autorecode_proc_func, &arc);
190 for (i = 0; i < arc.var_cnt; i++)
191 arc.dst_vars[i] = dict_create_var_assert (dataset_dict (ds),
192 arc.dst_names[i], 0);
196 return ok ? CMD_SUCCESS : CMD_CASCADING_FAILURE;
200 return CMD_CASCADING_FAILURE;
204 arc_free (struct autorecode_pgm *arc)
206 free (arc->src_vars);
207 if (arc->dst_names != NULL)
211 for (i = 0; i < arc->var_cnt; i++)
212 free (arc->dst_names[i]);
213 free (arc->dst_names);
215 free (arc->dst_vars);
216 if (arc->src_values != NULL)
220 for (i = 0; i < arc->var_cnt; i++)
221 hsh_destroy (arc->src_values[i]);
222 free (arc->src_values);
224 pool_destroy (arc->src_values_pool);
228 /* AUTORECODE transformation. */
231 recode (struct dataset *ds, const struct autorecode_pgm *arc)
233 struct autorecode_trns *trns;
236 trns = pool_create_container (struct autorecode_trns, pool);
237 trns->specs = pool_nalloc (trns->pool, arc->var_cnt, sizeof *trns->specs);
238 trns->spec_cnt = arc->var_cnt;
239 for (i = 0; i < arc->var_cnt; i++)
241 struct arc_spec *spec = &trns->specs[i];
242 void *const *p = hsh_sort (arc->src_values[i]);
243 int count = hsh_count (arc->src_values[i]);
246 spec->src = arc->src_vars[i];
247 spec->dest = arc->dst_vars[i];
249 if (var_is_alpha (arc->src_vars[i]))
250 spec->items = hsh_create (2 * count, compare_alpha_value,
251 hash_alpha_value, NULL, arc->src_vars[i]);
253 spec->items = hsh_create (2 * count, compare_numeric_value,
254 hash_numeric_value, NULL, NULL);
256 for (j = 0; *p; p++, j++)
258 struct arc_item *item = pool_alloc (trns->pool, sizeof *item);
259 union arc_value *vp = *p;
261 if (var_is_numeric (arc->src_vars[i]))
262 item->from.f = vp->f;
264 item->from.c = pool_clone (trns->pool, vp->c,
265 var_get_width (arc->src_vars[i]));
266 item->to = arc->direction == ASCENDING ? j + 1 : count - j;
267 hsh_force_insert (spec->items, item);
270 add_transformation (ds,
271 autorecode_trns_proc, autorecode_trns_free, trns);
274 /* Executes an AUTORECODE transformation. */
276 autorecode_trns_proc (void *trns_, struct ccase *c, casenumber case_idx UNUSED)
278 struct autorecode_trns *trns = trns_;
281 for (i = 0; i < trns->spec_cnt; i++)
283 struct arc_spec *spec = &trns->specs[i];
284 struct arc_item *item;
287 if (var_is_numeric (spec->src))
288 v.f = case_num (c, spec->src);
290 v.c = (char *) case_str (c, spec->src);
291 item = hsh_force_find (spec->items, &v);
293 case_data_rw (c, spec->dest)->f = item->to;
295 return TRNS_CONTINUE;
298 /* Frees an AUTORECODE transformation. */
300 autorecode_trns_free (void *trns_)
302 struct autorecode_trns *trns = trns_;
305 for (i = 0; i < trns->spec_cnt; i++)
306 hsh_destroy (trns->specs[i].items);
307 pool_destroy (trns->pool);
311 /* AUTORECODE procedure. */
314 compare_alpha_value (const void *a_, const void *b_, const void *v_)
316 const union arc_value *a = a_;
317 const union arc_value *b = b_;
318 const struct variable *v = v_;
320 return memcmp (a->c, b->c, var_get_width (v));
324 hash_alpha_value (const void *a_, const void *v_)
326 const union arc_value *a = a_;
327 const struct variable *v = v_;
329 return hsh_hash_bytes (a->c, var_get_width (v));
333 compare_numeric_value (const void *a_, const void *b_, const void *aux UNUSED)
335 const union arc_value *a = a_;
336 const union arc_value *b = b_;
338 return a->f < b->f ? -1 : a->f > b->f;
342 hash_numeric_value (const void *a_, const void *aux UNUSED)
344 const union arc_value *a = a_;
346 return hsh_hash_double (a->f);
350 autorecode_proc_func (const struct ccase *c, void *arc_, const struct dataset *ds UNUSED)
352 struct autorecode_pgm *arc = arc_;
355 for (i = 0; i < arc->var_cnt; i++)
357 union arc_value v, *vp, **vpp;
359 if (var_is_numeric (arc->src_vars[i]))
360 v.f = case_num (c, arc->src_vars[i]);
362 v.c = (char *) case_str (c, arc->src_vars[i]);
364 vpp = (union arc_value **) hsh_probe (arc->src_values[i], &v);
367 vp = pool_alloc (arc->src_values_pool, sizeof *vp);
368 if (var_is_numeric (arc->src_vars[i]))
371 vp->c = pool_clone (arc->src_values_pool,
372 v.c, var_get_width (arc->src_vars[i]));