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/variable-parser.h>
35 #include <language/lexer/range-parser.h>
36 #include <libpspp/alloc.h>
37 #include <libpspp/compiler.h>
38 #include <libpspp/magic.h>
39 #include <libpspp/message.h>
40 #include <libpspp/message.h>
41 #include <libpspp/pool.h>
42 #include <libpspp/str.h>
45 #define _(msgid) gettext (msgid)
49 /* Type of source value for RECODE. */
52 MAP_SINGLE, /* Specific value. */
53 MAP_RANGE, /* Range of values. */
54 MAP_SYSMIS, /* System missing value. */
55 MAP_MISSING, /* Any missing value. */
56 MAP_ELSE, /* Any value. */
57 MAP_CONVERT /* "123" => 123. */
60 /* A value involved in a RECODE mapping. */
63 double f; /* Numeric. */
64 char *c; /* Short or long string. */
67 /* Describes input values to be mapped. */
70 enum map_in_type type; /* One of MAP_*. */
71 union recode_value x, y; /* Source values. */
74 /* Describes the value used as output from a mapping. */
77 bool copy_input; /* If true, copy input to output. */
78 union recode_value value; /* If copy_input false, recoded value. */
79 int width; /* If copy_input false, output value width. */
82 /* Describes how to recode a single value or range of values into a
86 struct map_in in; /* Input values. */
87 struct map_out out; /* Output value. */
90 /* RECODE transformation. */
95 /* Variable types, for convenience. */
96 enum var_type src_type; /* src_vars[*]->type. */
97 enum var_type dst_type; /* dst_vars[*]->type. */
100 struct variable **src_vars; /* Source variables. */
101 struct variable **dst_vars; /* Destination variables. */
102 char **dst_names; /* Name of dest variables, if they're new. */
103 size_t var_cnt; /* Number of variables. */
106 struct mapping *mappings; /* Value mappings. */
107 size_t map_cnt; /* Number of mappings. */
110 static bool parse_src_vars (struct recode_trns *);
111 static bool parse_mappings (struct recode_trns *);
112 static bool parse_dst_vars (struct recode_trns *);
114 static void add_mapping (struct recode_trns *,
115 size_t *map_allocated, const struct map_in *);
117 static bool parse_map_in (struct map_in *, struct pool *,
118 enum var_type src_type, size_t max_src_width);
119 static void set_map_in_generic (struct map_in *, enum map_in_type);
120 static void set_map_in_num (struct map_in *, enum map_in_type, double, double);
121 static void set_map_in_str (struct map_in *, struct pool *,
122 const struct string *, size_t width);
124 static bool parse_map_out (struct pool *, struct map_out *);
125 static void set_map_out_num (struct map_out *, double);
126 static void set_map_out_str (struct map_out *, struct pool *,
127 const struct string *);
129 static void enlarge_dst_widths (struct recode_trns *);
130 static void create_dst_vars (struct recode_trns *);
132 static trns_proc_func recode_trns_proc;
133 static trns_free_func recode_trns_free;
137 /* Parses the RECODE transformation. */
143 struct recode_trns *trns
144 = pool_create_container (struct recode_trns, pool);
146 /* Parse source variable names,
147 then input to output mappings,
148 then destintation variable names. */
149 if (!parse_src_vars (trns)
150 || !parse_mappings (trns)
151 || !parse_dst_vars (trns))
153 recode_trns_free (trns);
157 /* Ensure that all the output strings are at least as wide
158 as the widest destination variable. */
159 if (trns->dst_type == ALPHA)
160 enlarge_dst_widths (trns);
162 /* Create destination variables, if needed.
163 This must be the final step; otherwise we'd have to
164 delete destination variables on failure. */
165 if (trns->src_vars != trns->dst_vars)
166 create_dst_vars (trns);
169 add_transformation (recode_trns_proc, recode_trns_free, trns);
171 while (lex_match ('/'));
173 return lex_end_of_command ();
176 /* Parses a set of variables to recode into TRNS->src_vars and
177 TRNS->var_cnt. Sets TRNS->src_type. Returns true if
178 successful, false on parse error. */
180 parse_src_vars (struct recode_trns *trns)
182 if (!parse_variables (default_dict, &trns->src_vars, &trns->var_cnt,
185 pool_register (trns->pool, free, trns->src_vars);
186 trns->src_type = trns->src_vars[0]->type;
190 /* Parses a set of mappings, which take the form (input=output),
191 into TRNS->mappings and TRNS->map_cnt. Sets TRNS->dst_type.
192 Returns true if successful, false on parse error. */
194 parse_mappings (struct recode_trns *trns)
196 size_t max_src_width;
197 size_t map_allocated;
201 /* Find length of longest source variable. */
202 max_src_width = trns->src_vars[0]->width;
203 for (i = 1; i < trns->var_cnt; i++)
205 size_t var_width = trns->src_vars[i]->width;
206 if (var_width > max_src_width)
207 max_src_width = var_width;
210 /* Parse the mappings in parentheses. */
211 trns->mappings = NULL;
214 have_dst_type = false;
215 if (!lex_force_match ('('))
219 enum var_type dst_type;
221 if (!lex_match_id ("CONVERT"))
224 size_t first_map_idx;
227 first_map_idx = trns->map_cnt;
229 /* Parse source specifications. */
233 if (!parse_map_in (&in, trns->pool,
234 trns->src_type, max_src_width))
236 add_mapping (trns, &map_allocated, &in);
239 while (!lex_match ('='));
241 if (!parse_map_out (trns->pool, &out))
243 dst_type = out.width == 0 ? NUMERIC : ALPHA;
244 if (have_dst_type && dst_type != trns->dst_type)
246 msg (SE, _("Inconsistent target variable types. "
248 "must be all numeric or all string."));
252 for (i = first_map_idx; i < trns->map_cnt; i++)
253 trns->mappings[i].out = out;
257 /* Parse CONVERT as a special case. */
259 set_map_in_generic (&in, MAP_CONVERT);
260 add_mapping (trns, &map_allocated, &in);
263 if (trns->src_type != ALPHA
264 || (have_dst_type && trns->dst_type != NUMERIC))
266 msg (SE, _("CONVERT requires string input values and "
267 "numeric output values."));
271 trns->dst_type = dst_type;
272 have_dst_type = true;
274 if (!lex_force_match (')'))
277 while (lex_match ('('));
282 /* Parses a mapping input value into IN, allocating memory from
283 POOL. The source value type must be provided as SRC_TYPE and,
284 if string, the maximum width of a string source variable must
285 be provided in MAX_SRC_WIDTH. Returns true if successful,
286 false on parse error. */
288 parse_map_in (struct map_in *in, struct pool *pool,
289 enum var_type src_type, size_t max_src_width)
291 if (lex_match_id ("ELSE"))
292 set_map_in_generic (in, MAP_ELSE);
293 else if (src_type == NUMERIC)
295 if (lex_match_id ("MISSING"))
296 set_map_in_generic (in, MAP_MISSING);
297 else if (lex_match_id ("SYSMIS"))
298 set_map_in_generic (in, MAP_SYSMIS);
302 if (!parse_num_range (&x, &y, NULL))
304 set_map_in_num (in, x == y ? MAP_SINGLE : MAP_RANGE, x, y);
309 if (!lex_force_string ())
311 set_map_in_str (in, pool, &tokstr, max_src_width);
318 /* Adds IN to the list of mappings in TRNS.
319 MAP_ALLOCATED is the current number of allocated mappings,
320 which is updated as needed. */
322 add_mapping (struct recode_trns *trns,
323 size_t *map_allocated, const struct map_in *in)
326 if (trns->map_cnt >= *map_allocated)
327 trns->mappings = pool_2nrealloc (trns->pool, trns->mappings,
329 sizeof *trns->mappings);
330 m = &trns->mappings[trns->map_cnt++];
334 /* Sets IN as a mapping of the given TYPE. */
336 set_map_in_generic (struct map_in *in, enum map_in_type type)
341 /* Sets IN as a numeric mapping of the given TYPE,
342 with X and Y as the two numeric values. */
344 set_map_in_num (struct map_in *in, enum map_in_type type, double x, double y)
351 /* Sets IN as a string mapping, with STRING as the string,
352 allocated from POOL. The string is padded with spaces on the
353 right to WIDTH characters long. */
355 set_map_in_str (struct map_in *in, struct pool *pool,
356 const struct string *string, size_t width)
358 in->type = MAP_SINGLE;
359 in->x.c = pool_alloc_unaligned (pool, width);
360 buf_copy_rpad (in->x.c, width, ds_data (string), ds_length (string));
363 /* Parses a mapping output value into OUT, allocating memory from
364 POOL. Returns true if successful, false on parse error. */
366 parse_map_out (struct pool *pool, struct map_out *out)
368 if (lex_is_number ())
370 set_map_out_num (out, lex_number ());
373 else if (lex_match_id ("SYSMIS"))
374 set_map_out_num (out, SYSMIS);
375 else if (token == T_STRING)
377 set_map_out_str (out, pool, &tokstr);
380 else if (lex_match_id ("COPY"))
381 out->copy_input = true;
384 lex_error (_("expecting output value"));
390 /* Sets OUT as a numeric mapping output with the given VALUE. */
392 set_map_out_num (struct map_out *out, double value)
394 out->copy_input = false;
395 out->value.f = value;
399 /* Sets OUT as a string mapping output with the given VALUE. */
401 set_map_out_str (struct map_out *out, struct pool *pool,
402 const struct string *value)
404 const char *string = ds_data (value);
405 size_t length = ds_length (value);
407 out->copy_input = false;
408 out->value.c = pool_alloc_unaligned (pool, length);
409 memcpy (out->value.c, string, length);
413 /* Parses a set of target variables into TRNS->dst_vars and
416 parse_dst_vars (struct recode_trns *trns)
420 if (lex_match_id ("INTO"))
425 if (!parse_mixed_vars_pool (trns->pool, &trns->dst_names, &name_cnt,
429 if (name_cnt != trns->var_cnt)
431 msg (SE, _("%u variable(s) cannot be recoded into "
432 "%u variable(s). Specify the same number "
433 "of variables as source and target variables."),
434 (unsigned) trns->var_cnt, (unsigned) name_cnt);
438 trns->dst_vars = pool_nalloc (trns->pool,
439 trns->var_cnt, sizeof *trns->dst_vars);
440 for (i = 0; i < trns->var_cnt; i++)
443 v = trns->dst_vars[i] = dict_lookup_var (default_dict,
445 if (v == NULL && trns->dst_type == ALPHA)
447 msg (SE, _("There is no variable named "
448 "%s. (All string variables specified "
449 "on INTO must already exist. Use the "
450 "STRING command to create a string "
459 trns->dst_vars = trns->src_vars;
460 if (trns->src_type != trns->dst_type)
462 msg (SE, _("INTO is required with %s input values "
463 "and %s output values."),
464 var_type_adj (trns->src_type),
465 var_type_adj (trns->dst_type));
470 for (i = 0; i < trns->var_cnt; i++)
472 struct variable *v = trns->dst_vars[i];
473 if (v != NULL && v->type != trns->dst_type)
475 msg (SE, _("Type mismatch. Cannot store %s data in "
477 trns->dst_type == ALPHA ? _("string") : _("numeric"),
478 v->type == ALPHA ? _("string") : _("numeric"),
487 /* Ensures that all the output values in TRNS are as wide as the
488 widest destination variable. */
490 enlarge_dst_widths (struct recode_trns *trns)
492 size_t max_dst_width;
496 for (i = 0; i < trns->var_cnt; i++)
498 struct variable *v = trns->dst_vars[i];
499 if (v->width > max_dst_width)
500 max_dst_width = v->width;
503 for (i = 0; i < trns->map_cnt; i++)
505 struct map_out *out = &trns->mappings[i].out;
506 if (!out->copy_input && out->width < max_dst_width)
508 char *s = pool_alloc_unaligned (trns->pool, max_dst_width + 1);
509 str_copy_rpad (s, max_dst_width + 1, out->value.c);
515 /* Creates destination variables that don't already exist. */
517 create_dst_vars (struct recode_trns *trns)
521 for (i = 0; i < trns->var_cnt; i++)
523 struct variable **var = &trns->dst_vars[i];
524 const char *name = trns->dst_names[i];
526 *var = dict_lookup_var (default_dict, name);
528 *var = dict_create_var_assert (default_dict, name, 0);
529 assert ((*var)->type == trns->dst_type);
533 /* Data transformation. */
535 /* Returns the output mapping in TRNS for an input of VALUE on
536 variable V, or a null pointer if there is no mapping. */
537 static const struct map_out *
538 find_src_numeric (struct recode_trns *trns, double value, struct variable *v)
542 for (m = trns->mappings; m < trns->mappings + trns->map_cnt; m++)
544 const struct map_in *in = &m->in;
545 const struct map_out *out = &m->out;
551 match = value == in->x.f;
554 match = mv_is_num_user_missing (&v->miss, value);
557 match = value >= in->x.f && value <= in->y.f;
573 /* Returns the output mapping in TRNS for an input of VALUE with
574 the given WIDTH, or a null pointer if there is no mapping. */
575 static const struct map_out *
576 find_src_string (struct recode_trns *trns, const char *value, int width)
580 for (m = trns->mappings; m < trns->mappings + trns->map_cnt; m++)
582 const struct map_in *in = &m->in;
583 struct map_out *out = &m->out;
589 match = !memcmp (value, in->x.c, width);
600 di.e = value + width;
602 di.flags = DI_IGNORE_ERROR;
604 di.format.type = FMT_F;
607 match = data_in (&di);
622 /* Performs RECODE transformation. */
624 recode_trns_proc (void *trns_, struct ccase *c, int case_idx UNUSED)
626 struct recode_trns *trns = trns_;
629 for (i = 0; i < trns->var_cnt; i++)
631 struct variable *src_var = trns->src_vars[i];
632 struct variable *dst_var = trns->dst_vars[i];
634 const union value *src_data = case_data (c, src_var->fv);
635 union value *dst_data = case_data_rw (c, dst_var->fv);
637 const struct map_out *out;
639 if (trns->src_type == NUMERIC)
640 out = find_src_numeric (trns, src_data->f, src_var);
642 out = find_src_string (trns, src_data->s, src_var->width);
644 if (trns->dst_type == NUMERIC)
647 dst_data->f = !out->copy_input ? out->value.f : src_data->f;
648 else if (trns->src_vars != trns->dst_vars)
649 dst_data->f = SYSMIS;
655 if (!out->copy_input)
656 memcpy (dst_data->s, out->value.c, dst_var->width);
657 else if (trns->src_vars != trns->dst_vars)
658 buf_copy_rpad (dst_data->s, dst_var->width,
659 src_data->s, src_var->width);
661 else if (trns->src_vars != trns->dst_vars)
662 memset (dst_data->s, ' ', dst_var->width);
666 return TRNS_CONTINUE;
669 /* Frees a RECODE transformation. */
671 recode_trns_free (void *trns_)
673 struct recode_trns *trns = trns_;
674 pool_destroy (trns->pool);