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
29 #include "dictionary.h"
34 #include "range-parser.h"
39 #define _(msgid) gettext (msgid)
43 /* Type of source value for RECODE. */
46 MAP_SINGLE, /* Specific value. */
47 MAP_RANGE, /* Range of values. */
48 MAP_SYSMIS, /* System missing value. */
49 MAP_MISSING, /* Any missing value. */
50 MAP_ELSE, /* Any value. */
51 MAP_CONVERT /* "123" => 123. */
54 /* Describes input values to be mapped. */
57 enum map_in_type type; /* One of MAP_*. */
58 union value x, y; /* Source values. */
61 /* Describes the value used as output from a mapping. */
64 bool copy_input; /* If true, copy input to output. */
65 union value value; /* If copy_input false, recoded value. */
66 int width; /* If copy_input false, output value width. */
69 /* Describes how to recode a single value or range of values into a
73 struct map_in in; /* Input values. */
74 struct map_out out; /* Output value. */
77 /* RECODE transformation. */
82 /* Variable types, for convenience. */
83 enum var_type src_type; /* src_vars[*]->type. */
84 enum var_type dst_type; /* dst_vars[*]->type. */
87 struct variable **src_vars; /* Source variables. */
88 struct variable **dst_vars; /* Destination variables. */
89 char **dst_names; /* Name of dest variables, if they're new. */
90 size_t var_cnt; /* Number of variables. */
93 struct mapping *mappings; /* Value mappings. */
94 size_t map_cnt; /* Number of mappings. */
97 static bool parse_src_vars (struct recode_trns *);
98 static bool parse_mappings (struct recode_trns *);
99 static bool parse_dst_vars (struct recode_trns *);
101 static void add_mapping (struct recode_trns *,
102 size_t *map_allocated, const struct map_in *);
104 static bool parse_map_in (struct map_in *, struct pool *,
105 enum var_type src_type, size_t max_src_width);
106 static void set_map_in_generic (struct map_in *, enum map_in_type);
107 static void set_map_in_num (struct map_in *, enum map_in_type, double, double);
108 static void set_map_in_str (struct map_in *, struct pool *,
109 const struct string *, size_t width);
111 static bool parse_map_out (struct pool *, struct map_out *);
112 static void set_map_out_num (struct map_out *, double);
113 static void set_map_out_str (struct map_out *, struct pool *,
114 const struct string *);
116 static void enlarge_dst_widths (struct recode_trns *);
117 static void create_dst_vars (struct recode_trns *);
119 static trns_proc_func recode_trns_proc;
120 static trns_free_func recode_trns_free;
124 /* Parses the RECODE transformation. */
130 struct recode_trns *trns
131 = pool_create_container (struct recode_trns, pool);
133 /* Parse source variable names,
134 then input to output mappings,
135 then destintation variable names. */
136 if (!parse_src_vars (trns)
137 || !parse_mappings (trns)
138 || !parse_dst_vars (trns))
140 recode_trns_free (trns);
141 return CMD_PART_SUCCESS;
144 /* Ensure that all the output strings are at least as wide
145 as the widest destination variable. */
146 if (trns->dst_type == ALPHA)
147 enlarge_dst_widths (trns);
149 /* Create destination variables, if needed.
150 This must be the final step; otherwise we'd have to
151 delete destination variables on failure. */
152 if (trns->src_vars != trns->dst_vars)
153 create_dst_vars (trns);
156 add_transformation (recode_trns_proc, recode_trns_free, trns);
158 while (lex_match ('/'));
160 return lex_end_of_command ();
163 /* Parses a set of variables to recode into TRNS->src_vars and
164 TRNS->var_cnt. Sets TRNS->src_type. Returns true if
165 successful, false on parse error. */
167 parse_src_vars (struct recode_trns *trns)
169 if (!parse_variables (default_dict, &trns->src_vars, &trns->var_cnt,
172 pool_register (trns->pool, free, trns->src_vars);
173 trns->src_type = trns->src_vars[0]->type;
177 /* Parses a set of mappings, which take the form (input=output),
178 into TRNS->mappings and TRNS->map_cnt. Sets TRNS->dst_type.
179 Returns true if successful, false on parse error. */
181 parse_mappings (struct recode_trns *trns)
183 size_t max_src_width;
184 size_t map_allocated;
188 /* Find length of longest source variable. */
189 max_src_width = trns->src_vars[0]->width;
190 for (i = 1; i < trns->var_cnt; i++)
192 size_t var_width = trns->src_vars[i]->width;
193 if (var_width > max_src_width)
194 max_src_width = var_width;
197 /* Parse the mappings in parentheses. */
198 trns->mappings = NULL;
201 have_dst_type = false;
202 if (!lex_force_match ('('))
206 enum var_type dst_type;
208 if (!lex_match_id ("CONVERT"))
211 size_t first_map_idx;
214 first_map_idx = trns->map_cnt;
216 /* Parse source specifications. */
220 if (!parse_map_in (&in, trns->pool,
221 trns->src_type, max_src_width))
223 add_mapping (trns, &map_allocated, &in);
226 while (!lex_match ('='));
228 if (!parse_map_out (trns->pool, &out))
230 dst_type = out.width == 0 ? NUMERIC : ALPHA;
231 if (have_dst_type && dst_type != trns->dst_type)
233 msg (SE, _("Inconsistent target variable types. "
235 "must be all numeric or all string."));
239 for (i = first_map_idx; i < trns->map_cnt; i++)
240 trns->mappings[i].out = out;
244 /* Parse CONVERT as a special case. */
246 set_map_in_generic (&in, MAP_CONVERT);
247 add_mapping (trns, &map_allocated, &in);
250 if (trns->src_type != ALPHA
251 || (have_dst_type && trns->dst_type != NUMERIC))
253 msg (SE, _("CONVERT requires string input values and "
254 "numeric output values."));
258 trns->dst_type = dst_type;
259 have_dst_type = true;
261 if (!lex_force_match (')'))
264 while (lex_match ('('));
269 /* Parses a mapping input value into IN, allocating memory from
270 POOL. The source value type must be provided as SRC_TYPE and,
271 if string, the maximum width of a string source variable must
272 be provided in MAX_SRC_WIDTH. Returns true if successful,
273 false on parse error. */
275 parse_map_in (struct map_in *in, struct pool *pool,
276 enum var_type src_type, size_t max_src_width)
278 if (lex_match_id ("ELSE"))
279 set_map_in_generic (in, MAP_ELSE);
280 else if (src_type == NUMERIC)
282 if (lex_match_id ("MISSING"))
283 set_map_in_generic (in, MAP_MISSING);
284 else if (lex_match_id ("SYSMIS"))
285 set_map_in_generic (in, MAP_SYSMIS);
289 if (!parse_num_range (&x, &y, NULL))
291 set_map_in_num (in, x == y ? MAP_SINGLE : MAP_RANGE, x, y);
296 if (!lex_force_string ())
298 set_map_in_str (in, pool, &tokstr, max_src_width);
305 /* Adds IN to the list of mappings in TRNS.
306 MAP_ALLOCATED is the current number of allocated mappings,
307 which is updated as needed. */
309 add_mapping (struct recode_trns *trns,
310 size_t *map_allocated, const struct map_in *in)
313 if (trns->map_cnt >= *map_allocated)
314 trns->mappings = pool_2nrealloc (trns->pool, trns->mappings,
316 sizeof *trns->mappings);
317 m = &trns->mappings[trns->map_cnt++];
321 /* Sets IN as a mapping of the given TYPE. */
323 set_map_in_generic (struct map_in *in, enum map_in_type type)
328 /* Sets IN as a numeric mapping of the given TYPE,
329 with X and Y as the two numeric values. */
331 set_map_in_num (struct map_in *in, enum map_in_type type, double x, double y)
338 /* Sets IN as a string mapping, with STRING as the string,
339 allocated from POOL. The string is padded with spaces on the
340 right to WIDTH characters long. */
342 set_map_in_str (struct map_in *in, struct pool *pool,
343 const struct string *string, size_t width)
345 in->type = MAP_SINGLE;
346 in->x.c = pool_alloc_unaligned (pool, width);
347 buf_copy_rpad (in->x.c, width, ds_data (string), ds_length (string));
350 /* Parses a mapping output value into OUT, allocating memory from
351 POOL. Returns true if successful, false on parse error. */
353 parse_map_out (struct pool *pool, struct map_out *out)
355 if (lex_is_number ())
357 set_map_out_num (out, lex_number ());
360 else if (lex_match_id ("SYSMIS"))
361 set_map_out_num (out, SYSMIS);
362 else if (token == T_STRING)
364 set_map_out_str (out, pool, &tokstr);
367 else if (lex_match_id ("COPY"))
368 out->copy_input = true;
371 lex_error (_("expecting output value"));
377 /* Sets OUT as a numeric mapping output with the given VALUE. */
379 set_map_out_num (struct map_out *out, double value)
381 out->copy_input = false;
382 out->value.f = value;
386 /* Sets OUT as a string mapping output with the given VALUE. */
388 set_map_out_str (struct map_out *out, struct pool *pool,
389 const struct string *value)
391 const char *string = ds_data (value);
392 size_t length = ds_length (value);
394 out->copy_input = false;
395 out->value.c = pool_alloc_unaligned (pool, length);
396 memcpy (out->value.c, string, length);
400 /* Parses a set of target variables into TRNS->dst_vars and
403 parse_dst_vars (struct recode_trns *trns)
407 if (lex_match_id ("INTO"))
412 if (!parse_mixed_vars_pool (trns->pool, &trns->dst_names, &name_cnt,
416 if (name_cnt != trns->var_cnt)
418 msg (SE, _("%u variable(s) cannot be recoded into "
419 "%u variable(s). Specify the same number "
420 "of variables as source and target variables."),
421 (unsigned) trns->var_cnt, (unsigned) name_cnt);
425 trns->dst_vars = pool_nalloc (trns->pool,
426 trns->var_cnt, sizeof *trns->dst_vars);
427 for (i = 0; i < trns->var_cnt; i++)
430 v = trns->dst_vars[i] = dict_lookup_var (default_dict,
432 if (v == NULL && trns->dst_type == ALPHA)
434 msg (SE, _("There is no variable named "
435 "%s. (All string variables specified "
436 "on INTO must already exist. Use the "
437 "STRING command to create a string "
446 trns->dst_vars = trns->src_vars;
447 if (trns->src_type != trns->dst_type)
449 msg (SE, _("INTO is required with %s input values "
450 "and %s output values."),
451 var_type_adj (trns->src_type),
452 var_type_adj (trns->dst_type));
457 for (i = 0; i < trns->var_cnt; i++)
459 struct variable *v = trns->dst_vars[i];
460 if (v != NULL && v->type != trns->dst_type)
462 msg (SE, _("Type mismatch. Cannot store %s data in "
464 trns->dst_type == ALPHA ? _("string") : _("numeric"),
465 v->type == ALPHA ? _("string") : _("numeric"),
474 /* Ensures that all the output values in TRNS are as wide as the
475 widest destination variable. */
477 enlarge_dst_widths (struct recode_trns *trns)
479 size_t max_dst_width;
483 for (i = 0; i < trns->var_cnt; i++)
485 struct variable *v = trns->dst_vars[i];
486 if (v->width > max_dst_width)
487 max_dst_width = v->width;
490 for (i = 0; i < trns->map_cnt; i++)
492 struct map_out *out = &trns->mappings[i].out;
493 if (!out->copy_input && out->width < max_dst_width)
495 char *s = pool_alloc_unaligned (trns->pool, max_dst_width + 1);
496 str_copy_rpad (s, max_dst_width + 1, out->value.c);
502 /* Creates destination variables that don't already exist. */
504 create_dst_vars (struct recode_trns *trns)
508 for (i = 0; i < trns->var_cnt; i++)
510 struct variable **var = &trns->dst_vars[i];
511 const char *name = trns->dst_names[i];
513 *var = dict_lookup_var (default_dict, name);
515 *var = dict_create_var_assert (default_dict, name, 0);
516 assert ((*var)->type == trns->dst_type);
520 /* Data transformation. */
522 /* Returns the output mapping in TRNS for an input of VALUE on
523 variable V, or a null pointer if there is no mapping. */
524 static const struct map_out *
525 find_src_numeric (struct recode_trns *trns, double value, struct variable *v)
529 for (m = trns->mappings; m < trns->mappings + trns->map_cnt; m++)
531 const struct map_in *in = &m->in;
532 const struct map_out *out = &m->out;
538 match = value == in->x.f;
541 match = mv_is_num_user_missing (&v->miss, value);
544 match = value >= in->x.f && value <= in->y.f;
560 /* Returns the output mapping in TRNS for an input of VALUE with
561 the given WIDTH, or a null pointer if there is no mapping. */
562 static const struct map_out *
563 find_src_string (struct recode_trns *trns, const char *value, int width)
567 for (m = trns->mappings; m < trns->mappings + trns->map_cnt; m++)
569 const struct map_in *in = &m->in;
570 struct map_out *out = &m->out;
576 match = !memcmp (value, in->x.c, width);
586 di.e = value + width;
588 di.flags = DI_IGNORE_ERROR;
590 di.format.type = FMT_F;
593 match = data_in (&di);
607 /* Performs RECODE transformation. */
609 recode_trns_proc (void *trns_, struct ccase *c, int case_idx UNUSED)
611 struct recode_trns *trns = trns_;
614 for (i = 0; i < trns->var_cnt; i++)
616 struct variable *src_var = trns->src_vars[i];
617 struct variable *dst_var = trns->dst_vars[i];
619 const union value *src_data = case_data (c, src_var->fv);
620 union value *dst_data = case_data_rw (c, dst_var->fv);
622 const struct map_out *out;
624 if (trns->src_type == NUMERIC)
625 out = find_src_numeric (trns, src_data->f, src_var);
627 out = find_src_string (trns, src_data->s, src_var->width);
629 if (trns->dst_type == NUMERIC)
632 dst_data->f = !out->copy_input ? out->value.f : src_data->f;
633 else if (trns->src_vars != trns->dst_vars)
634 dst_data->f = SYSMIS;
640 if (!out->copy_input)
641 memcpy (dst_data->s, out->value.c, dst_var->width);
642 else if (trns->src_vars != trns->dst_vars)
643 buf_copy_rpad (dst_data->s, dst_var->width,
644 src_data->s, src_var->width);
646 else if (trns->src_vars != trns->dst_vars)
647 memset (dst_data->s, ' ', dst_var->width);
651 return TRNS_CONTINUE;
654 /* Frees a RECODE transformation. */
656 recode_trns_free (void *trns_)
658 struct recode_trns *trns = trns_;
659 pool_destroy (trns->pool);