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 <data/case.h>
27 #include <data/data-in.h>
28 #include <data/dictionary.h>
29 #include <data/procedure.h>
30 #include <data/transformations.h>
31 #include <data/variable.h>
32 #include <language/command.h>
33 #include <language/lexer/lexer.h>
34 #include <language/lexer/range-parser.h>
35 #include <libpspp/alloc.h>
36 #include <libpspp/compiler.h>
37 #include <libpspp/magic.h>
38 #include <libpspp/message.h>
39 #include <libpspp/message.h>
40 #include <libpspp/pool.h>
41 #include <libpspp/str.h>
44 #define _(msgid) gettext (msgid)
48 /* Type of source value for RECODE. */
51 MAP_SINGLE, /* Specific value. */
52 MAP_RANGE, /* Range of values. */
53 MAP_SYSMIS, /* System missing value. */
54 MAP_MISSING, /* Any missing value. */
55 MAP_ELSE, /* Any value. */
56 MAP_CONVERT /* "123" => 123. */
59 /* Describes input values to be mapped. */
62 enum map_in_type type; /* One of MAP_*. */
63 union value x, y; /* Source values. */
66 /* Describes the value used as output from a mapping. */
69 bool copy_input; /* If true, copy input to output. */
70 union value value; /* If copy_input false, recoded value. */
71 int width; /* If copy_input false, output value width. */
74 /* Describes how to recode a single value or range of values into a
78 struct map_in in; /* Input values. */
79 struct map_out out; /* Output value. */
82 /* RECODE transformation. */
87 /* Variable types, for convenience. */
88 enum var_type src_type; /* src_vars[*]->type. */
89 enum var_type dst_type; /* dst_vars[*]->type. */
92 struct variable **src_vars; /* Source variables. */
93 struct variable **dst_vars; /* Destination variables. */
94 char **dst_names; /* Name of dest variables, if they're new. */
95 size_t var_cnt; /* Number of variables. */
98 struct mapping *mappings; /* Value mappings. */
99 size_t map_cnt; /* Number of mappings. */
102 static bool parse_src_vars (struct recode_trns *);
103 static bool parse_mappings (struct recode_trns *);
104 static bool parse_dst_vars (struct recode_trns *);
106 static void add_mapping (struct recode_trns *,
107 size_t *map_allocated, const struct map_in *);
109 static bool parse_map_in (struct map_in *, struct pool *,
110 enum var_type src_type, size_t max_src_width);
111 static void set_map_in_generic (struct map_in *, enum map_in_type);
112 static void set_map_in_num (struct map_in *, enum map_in_type, double, double);
113 static void set_map_in_str (struct map_in *, struct pool *,
114 const struct string *, size_t width);
116 static bool parse_map_out (struct pool *, struct map_out *);
117 static void set_map_out_num (struct map_out *, double);
118 static void set_map_out_str (struct map_out *, struct pool *,
119 const struct string *);
121 static void enlarge_dst_widths (struct recode_trns *);
122 static void create_dst_vars (struct recode_trns *);
124 static trns_proc_func recode_trns_proc;
125 static trns_free_func recode_trns_free;
129 /* Parses the RECODE transformation. */
135 struct recode_trns *trns
136 = pool_create_container (struct recode_trns, pool);
138 /* Parse source variable names,
139 then input to output mappings,
140 then destintation variable names. */
141 if (!parse_src_vars (trns)
142 || !parse_mappings (trns)
143 || !parse_dst_vars (trns))
145 recode_trns_free (trns);
149 /* Ensure that all the output strings are at least as wide
150 as the widest destination variable. */
151 if (trns->dst_type == ALPHA)
152 enlarge_dst_widths (trns);
154 /* Create destination variables, if needed.
155 This must be the final step; otherwise we'd have to
156 delete destination variables on failure. */
157 if (trns->src_vars != trns->dst_vars)
158 create_dst_vars (trns);
161 add_transformation (recode_trns_proc, recode_trns_free, trns);
163 while (lex_match ('/'));
165 return lex_end_of_command ();
168 /* Parses a set of variables to recode into TRNS->src_vars and
169 TRNS->var_cnt. Sets TRNS->src_type. Returns true if
170 successful, false on parse error. */
172 parse_src_vars (struct recode_trns *trns)
174 if (!parse_variables (default_dict, &trns->src_vars, &trns->var_cnt,
177 pool_register (trns->pool, free, trns->src_vars);
178 trns->src_type = trns->src_vars[0]->type;
182 /* Parses a set of mappings, which take the form (input=output),
183 into TRNS->mappings and TRNS->map_cnt. Sets TRNS->dst_type.
184 Returns true if successful, false on parse error. */
186 parse_mappings (struct recode_trns *trns)
188 size_t max_src_width;
189 size_t map_allocated;
193 /* Find length of longest source variable. */
194 max_src_width = trns->src_vars[0]->width;
195 for (i = 1; i < trns->var_cnt; i++)
197 size_t var_width = trns->src_vars[i]->width;
198 if (var_width > max_src_width)
199 max_src_width = var_width;
202 /* Parse the mappings in parentheses. */
203 trns->mappings = NULL;
206 have_dst_type = false;
207 if (!lex_force_match ('('))
211 enum var_type dst_type;
213 if (!lex_match_id ("CONVERT"))
216 size_t first_map_idx;
219 first_map_idx = trns->map_cnt;
221 /* Parse source specifications. */
225 if (!parse_map_in (&in, trns->pool,
226 trns->src_type, max_src_width))
228 add_mapping (trns, &map_allocated, &in);
231 while (!lex_match ('='));
233 if (!parse_map_out (trns->pool, &out))
235 dst_type = out.width == 0 ? NUMERIC : ALPHA;
236 if (have_dst_type && dst_type != trns->dst_type)
238 msg (SE, _("Inconsistent target variable types. "
240 "must be all numeric or all string."));
244 for (i = first_map_idx; i < trns->map_cnt; i++)
245 trns->mappings[i].out = out;
249 /* Parse CONVERT as a special case. */
251 set_map_in_generic (&in, MAP_CONVERT);
252 add_mapping (trns, &map_allocated, &in);
255 if (trns->src_type != ALPHA
256 || (have_dst_type && trns->dst_type != NUMERIC))
258 msg (SE, _("CONVERT requires string input values and "
259 "numeric output values."));
263 trns->dst_type = dst_type;
264 have_dst_type = true;
266 if (!lex_force_match (')'))
269 while (lex_match ('('));
274 /* Parses a mapping input value into IN, allocating memory from
275 POOL. The source value type must be provided as SRC_TYPE and,
276 if string, the maximum width of a string source variable must
277 be provided in MAX_SRC_WIDTH. Returns true if successful,
278 false on parse error. */
280 parse_map_in (struct map_in *in, struct pool *pool,
281 enum var_type src_type, size_t max_src_width)
283 if (lex_match_id ("ELSE"))
284 set_map_in_generic (in, MAP_ELSE);
285 else if (src_type == NUMERIC)
287 if (lex_match_id ("MISSING"))
288 set_map_in_generic (in, MAP_MISSING);
289 else if (lex_match_id ("SYSMIS"))
290 set_map_in_generic (in, MAP_SYSMIS);
294 if (!parse_num_range (&x, &y, NULL))
296 set_map_in_num (in, x == y ? MAP_SINGLE : MAP_RANGE, x, y);
301 if (!lex_force_string ())
303 set_map_in_str (in, pool, &tokstr, max_src_width);
310 /* Adds IN to the list of mappings in TRNS.
311 MAP_ALLOCATED is the current number of allocated mappings,
312 which is updated as needed. */
314 add_mapping (struct recode_trns *trns,
315 size_t *map_allocated, const struct map_in *in)
318 if (trns->map_cnt >= *map_allocated)
319 trns->mappings = pool_2nrealloc (trns->pool, trns->mappings,
321 sizeof *trns->mappings);
322 m = &trns->mappings[trns->map_cnt++];
326 /* Sets IN as a mapping of the given TYPE. */
328 set_map_in_generic (struct map_in *in, enum map_in_type type)
333 /* Sets IN as a numeric mapping of the given TYPE,
334 with X and Y as the two numeric values. */
336 set_map_in_num (struct map_in *in, enum map_in_type type, double x, double y)
343 /* Sets IN as a string mapping, with STRING as the string,
344 allocated from POOL. The string is padded with spaces on the
345 right to WIDTH characters long. */
347 set_map_in_str (struct map_in *in, struct pool *pool,
348 const struct string *string, size_t width)
350 in->type = MAP_SINGLE;
351 in->x.c = pool_alloc_unaligned (pool, width);
352 buf_copy_rpad (in->x.c, width, ds_data (string), ds_length (string));
355 /* Parses a mapping output value into OUT, allocating memory from
356 POOL. Returns true if successful, false on parse error. */
358 parse_map_out (struct pool *pool, struct map_out *out)
360 if (lex_is_number ())
362 set_map_out_num (out, lex_number ());
365 else if (lex_match_id ("SYSMIS"))
366 set_map_out_num (out, SYSMIS);
367 else if (token == T_STRING)
369 set_map_out_str (out, pool, &tokstr);
372 else if (lex_match_id ("COPY"))
373 out->copy_input = true;
376 lex_error (_("expecting output value"));
382 /* Sets OUT as a numeric mapping output with the given VALUE. */
384 set_map_out_num (struct map_out *out, double value)
386 out->copy_input = false;
387 out->value.f = value;
391 /* Sets OUT as a string mapping output with the given VALUE. */
393 set_map_out_str (struct map_out *out, struct pool *pool,
394 const struct string *value)
396 const char *string = ds_data (value);
397 size_t length = ds_length (value);
399 out->copy_input = false;
400 out->value.c = pool_alloc_unaligned (pool, length);
401 memcpy (out->value.c, string, length);
405 /* Parses a set of target variables into TRNS->dst_vars and
408 parse_dst_vars (struct recode_trns *trns)
412 if (lex_match_id ("INTO"))
417 if (!parse_mixed_vars_pool (trns->pool, &trns->dst_names, &name_cnt,
421 if (name_cnt != trns->var_cnt)
423 msg (SE, _("%u variable(s) cannot be recoded into "
424 "%u variable(s). Specify the same number "
425 "of variables as source and target variables."),
426 (unsigned) trns->var_cnt, (unsigned) name_cnt);
430 trns->dst_vars = pool_nalloc (trns->pool,
431 trns->var_cnt, sizeof *trns->dst_vars);
432 for (i = 0; i < trns->var_cnt; i++)
435 v = trns->dst_vars[i] = dict_lookup_var (default_dict,
437 if (v == NULL && trns->dst_type == ALPHA)
439 msg (SE, _("There is no variable named "
440 "%s. (All string variables specified "
441 "on INTO must already exist. Use the "
442 "STRING command to create a string "
451 trns->dst_vars = trns->src_vars;
452 if (trns->src_type != trns->dst_type)
454 msg (SE, _("INTO is required with %s input values "
455 "and %s output values."),
456 var_type_adj (trns->src_type),
457 var_type_adj (trns->dst_type));
462 for (i = 0; i < trns->var_cnt; i++)
464 struct variable *v = trns->dst_vars[i];
465 if (v != NULL && v->type != trns->dst_type)
467 msg (SE, _("Type mismatch. Cannot store %s data in "
469 trns->dst_type == ALPHA ? _("string") : _("numeric"),
470 v->type == ALPHA ? _("string") : _("numeric"),
479 /* Ensures that all the output values in TRNS are as wide as the
480 widest destination variable. */
482 enlarge_dst_widths (struct recode_trns *trns)
484 size_t max_dst_width;
488 for (i = 0; i < trns->var_cnt; i++)
490 struct variable *v = trns->dst_vars[i];
491 if (v->width > max_dst_width)
492 max_dst_width = v->width;
495 for (i = 0; i < trns->map_cnt; i++)
497 struct map_out *out = &trns->mappings[i].out;
498 if (!out->copy_input && out->width < max_dst_width)
500 char *s = pool_alloc_unaligned (trns->pool, max_dst_width + 1);
501 str_copy_rpad (s, max_dst_width + 1, out->value.c);
507 /* Creates destination variables that don't already exist. */
509 create_dst_vars (struct recode_trns *trns)
513 for (i = 0; i < trns->var_cnt; i++)
515 struct variable **var = &trns->dst_vars[i];
516 const char *name = trns->dst_names[i];
518 *var = dict_lookup_var (default_dict, name);
520 *var = dict_create_var_assert (default_dict, name, 0);
521 assert ((*var)->type == trns->dst_type);
525 /* Data transformation. */
527 /* Returns the output mapping in TRNS for an input of VALUE on
528 variable V, or a null pointer if there is no mapping. */
529 static const struct map_out *
530 find_src_numeric (struct recode_trns *trns, double value, struct variable *v)
534 for (m = trns->mappings; m < trns->mappings + trns->map_cnt; m++)
536 const struct map_in *in = &m->in;
537 const struct map_out *out = &m->out;
543 match = value == in->x.f;
546 match = mv_is_num_user_missing (&v->miss, value);
549 match = value >= in->x.f && value <= in->y.f;
565 /* Returns the output mapping in TRNS for an input of VALUE with
566 the given WIDTH, or a null pointer if there is no mapping. */
567 static const struct map_out *
568 find_src_string (struct recode_trns *trns, const char *value, int width)
572 for (m = trns->mappings; m < trns->mappings + trns->map_cnt; m++)
574 const struct map_in *in = &m->in;
575 struct map_out *out = &m->out;
581 match = !memcmp (value, in->x.c, width);
591 di.e = value + width;
593 di.flags = DI_IGNORE_ERROR;
595 di.format.type = FMT_F;
598 match = data_in (&di);
612 /* Performs RECODE transformation. */
614 recode_trns_proc (void *trns_, struct ccase *c, int case_idx UNUSED)
616 struct recode_trns *trns = trns_;
619 for (i = 0; i < trns->var_cnt; i++)
621 struct variable *src_var = trns->src_vars[i];
622 struct variable *dst_var = trns->dst_vars[i];
624 const union value *src_data = case_data (c, src_var->fv);
625 union value *dst_data = case_data_rw (c, dst_var->fv);
627 const struct map_out *out;
629 if (trns->src_type == NUMERIC)
630 out = find_src_numeric (trns, src_data->f, src_var);
632 out = find_src_string (trns, src_data->s, src_var->width);
634 if (trns->dst_type == NUMERIC)
637 dst_data->f = !out->copy_input ? out->value.f : src_data->f;
638 else if (trns->src_vars != trns->dst_vars)
639 dst_data->f = SYSMIS;
645 if (!out->copy_input)
646 memcpy (dst_data->s, out->value.c, dst_var->width);
647 else if (trns->src_vars != trns->dst_vars)
648 buf_copy_rpad (dst_data->s, dst_var->width,
649 src_data->s, src_var->width);
651 else if (trns->src_vars != trns->dst_vars)
652 memset (dst_data->s, ' ', dst_var->width);
656 return TRNS_CONTINUE;
659 /* Frees a RECODE transformation. */
661 recode_trns_free (void *trns_)
663 struct recode_trns *trns = trns_;
664 pool_destroy (trns->pool);