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/assertion.h>
38 #include <libpspp/compiler.h>
39 #include <libpspp/magic.h>
40 #include <libpspp/message.h>
41 #include <libpspp/message.h>
42 #include <libpspp/pool.h>
43 #include <libpspp/str.h>
46 #define _(msgid) gettext (msgid)
50 /* Type of source value for RECODE. */
53 MAP_SINGLE, /* Specific value. */
54 MAP_RANGE, /* Range of values. */
55 MAP_SYSMIS, /* System missing value. */
56 MAP_MISSING, /* Any missing value. */
57 MAP_ELSE, /* Any value. */
58 MAP_CONVERT /* "123" => 123. */
61 /* A value involved in a RECODE mapping. */
64 double f; /* Numeric. */
65 char *c; /* Short or long string. */
68 /* Describes input values to be mapped. */
71 enum map_in_type type; /* One of MAP_*. */
72 union recode_value x, y; /* Source values. */
75 /* Describes the value used as output from a mapping. */
78 bool copy_input; /* If true, copy input to output. */
79 union recode_value value; /* If copy_input false, recoded value. */
80 int width; /* If copy_input false, output value width. */
83 /* Describes how to recode a single value or range of values into a
87 struct map_in in; /* Input values. */
88 struct map_out out; /* Output value. */
91 /* RECODE transformation. */
96 /* Variable types, for convenience. */
97 enum var_type src_type; /* src_vars[*]->type. */
98 enum var_type dst_type; /* dst_vars[*]->type. */
101 struct variable **src_vars; /* Source variables. */
102 struct variable **dst_vars; /* Destination variables. */
103 char **dst_names; /* Name of dest variables, if they're new. */
104 size_t var_cnt; /* Number of variables. */
107 struct mapping *mappings; /* Value mappings. */
108 size_t map_cnt; /* Number of mappings. */
111 static bool parse_src_vars (struct lexer *, struct recode_trns *, const struct dictionary *dict);
112 static bool parse_mappings (struct lexer *, struct recode_trns *);
113 static bool parse_dst_vars (struct lexer *, struct recode_trns *, const struct dictionary *dict);
115 static void add_mapping (struct recode_trns *,
116 size_t *map_allocated, const struct map_in *);
118 static bool parse_map_in (struct lexer *lexer, struct map_in *, struct pool *,
119 enum var_type src_type, size_t max_src_width);
120 static void set_map_in_generic (struct map_in *, enum map_in_type);
121 static void set_map_in_num (struct map_in *, enum map_in_type, double, double);
122 static void set_map_in_str (struct map_in *, struct pool *,
123 const struct string *, size_t width);
125 static bool parse_map_out (struct lexer *lexer, struct pool *, struct map_out *);
126 static void set_map_out_num (struct map_out *, double);
127 static void set_map_out_str (struct map_out *, struct pool *,
128 const struct string *);
130 static void enlarge_dst_widths (struct recode_trns *);
131 static void create_dst_vars (struct recode_trns *, struct dictionary *);
133 static trns_proc_func recode_trns_proc;
134 static trns_free_func recode_trns_free;
138 /* Parses the RECODE transformation. */
140 cmd_recode (struct lexer *lexer, struct dataset *ds)
144 struct recode_trns *trns
145 = pool_create_container (struct recode_trns, pool);
147 /* Parse source variable names,
148 then input to output mappings,
149 then destintation variable names. */
150 if (!parse_src_vars (lexer, trns, dataset_dict (ds) )
151 || !parse_mappings (lexer, trns)
152 || !parse_dst_vars (lexer, trns, dataset_dict (ds)))
154 recode_trns_free (trns);
158 /* Ensure that all the output strings are at least as wide
159 as the widest destination variable. */
160 if (trns->dst_type == ALPHA)
161 enlarge_dst_widths (trns);
163 /* Create destination variables, if needed.
164 This must be the final step; otherwise we'd have to
165 delete destination variables on failure. */
166 if (trns->src_vars != trns->dst_vars)
167 create_dst_vars (trns, dataset_dict (ds));
170 add_transformation (ds,
171 recode_trns_proc, recode_trns_free, trns);
173 while (lex_match (lexer, '/'));
175 return lex_end_of_command (lexer);
178 /* Parses a set of variables to recode into TRNS->src_vars and
179 TRNS->var_cnt. Sets TRNS->src_type. Returns true if
180 successful, false on parse error. */
182 parse_src_vars (struct lexer *lexer,
183 struct recode_trns *trns, const struct dictionary *dict)
185 if (!parse_variables (lexer, dict, &trns->src_vars, &trns->var_cnt,
188 pool_register (trns->pool, free, trns->src_vars);
189 trns->src_type = trns->src_vars[0]->type;
193 /* Parses a set of mappings, which take the form (input=output),
194 into TRNS->mappings and TRNS->map_cnt. Sets TRNS->dst_type.
195 Returns true if successful, false on parse error. */
197 parse_mappings (struct lexer *lexer, struct recode_trns *trns)
199 size_t max_src_width;
200 size_t map_allocated;
204 /* Find length of longest source variable. */
205 max_src_width = trns->src_vars[0]->width;
206 for (i = 1; i < trns->var_cnt; i++)
208 size_t var_width = trns->src_vars[i]->width;
209 if (var_width > max_src_width)
210 max_src_width = var_width;
213 /* Parse the mappings in parentheses. */
214 trns->mappings = NULL;
217 have_dst_type = false;
218 if (!lex_force_match (lexer, '('))
222 enum var_type dst_type;
224 if (!lex_match_id (lexer, "CONVERT"))
227 size_t first_map_idx;
230 first_map_idx = trns->map_cnt;
232 /* Parse source specifications. */
236 if (!parse_map_in (lexer, &in, trns->pool,
237 trns->src_type, max_src_width))
239 add_mapping (trns, &map_allocated, &in);
240 lex_match (lexer, ',');
242 while (!lex_match (lexer, '='));
244 if (!parse_map_out (lexer, trns->pool, &out))
246 dst_type = out.width == 0 ? NUMERIC : ALPHA;
247 if (have_dst_type && dst_type != trns->dst_type)
249 msg (SE, _("Inconsistent target variable types. "
251 "must be all numeric or all string."));
255 for (i = first_map_idx; i < trns->map_cnt; i++)
256 trns->mappings[i].out = out;
260 /* Parse CONVERT as a special case. */
262 set_map_in_generic (&in, MAP_CONVERT);
263 add_mapping (trns, &map_allocated, &in);
266 if (trns->src_type != ALPHA
267 || (have_dst_type && trns->dst_type != NUMERIC))
269 msg (SE, _("CONVERT requires string input values and "
270 "numeric output values."));
274 trns->dst_type = dst_type;
275 have_dst_type = true;
277 if (!lex_force_match (lexer, ')'))
280 while (lex_match (lexer, '('));
285 /* Parses a mapping input value into IN, allocating memory from
286 POOL. The source value type must be provided as SRC_TYPE and,
287 if string, the maximum width of a string source variable must
288 be provided in MAX_SRC_WIDTH. Returns true if successful,
289 false on parse error. */
291 parse_map_in (struct lexer *lexer, struct map_in *in, struct pool *pool,
292 enum var_type src_type, size_t max_src_width)
294 if (lex_match_id (lexer, "ELSE"))
295 set_map_in_generic (in, MAP_ELSE);
296 else if (src_type == NUMERIC)
298 if (lex_match_id (lexer, "MISSING"))
299 set_map_in_generic (in, MAP_MISSING);
300 else if (lex_match_id (lexer, "SYSMIS"))
301 set_map_in_generic (in, MAP_SYSMIS);
305 if (!parse_num_range (lexer, &x, &y, NULL))
307 set_map_in_num (in, x == y ? MAP_SINGLE : MAP_RANGE, x, y);
312 if (!lex_force_string (lexer))
314 set_map_in_str (in, pool, lex_tokstr (lexer), max_src_width);
321 /* Adds IN to the list of mappings in TRNS.
322 MAP_ALLOCATED is the current number of allocated mappings,
323 which is updated as needed. */
325 add_mapping (struct recode_trns *trns,
326 size_t *map_allocated, const struct map_in *in)
329 if (trns->map_cnt >= *map_allocated)
330 trns->mappings = pool_2nrealloc (trns->pool, trns->mappings,
332 sizeof *trns->mappings);
333 m = &trns->mappings[trns->map_cnt++];
337 /* Sets IN as a mapping of the given TYPE. */
339 set_map_in_generic (struct map_in *in, enum map_in_type type)
344 /* Sets IN as a numeric mapping of the given TYPE,
345 with X and Y as the two numeric values. */
347 set_map_in_num (struct map_in *in, enum map_in_type type, double x, double y)
354 /* Sets IN as a string mapping, with STRING as the string,
355 allocated from POOL. The string is padded with spaces on the
356 right to WIDTH characters long. */
358 set_map_in_str (struct map_in *in, struct pool *pool,
359 const struct string *string, size_t width)
361 in->type = MAP_SINGLE;
362 in->x.c = pool_alloc_unaligned (pool, width);
363 buf_copy_rpad (in->x.c, width, ds_data (string), ds_length (string));
366 /* Parses a mapping output value into OUT, allocating memory from
367 POOL. Returns true if successful, false on parse error. */
369 parse_map_out (struct lexer *lexer, struct pool *pool, struct map_out *out)
371 if (lex_is_number (lexer))
373 set_map_out_num (out, lex_number (lexer));
376 else if (lex_match_id (lexer, "SYSMIS"))
377 set_map_out_num (out, SYSMIS);
378 else if (lex_token (lexer) == T_STRING)
380 set_map_out_str (out, pool, lex_tokstr (lexer));
383 else if (lex_match_id (lexer, "COPY"))
384 out->copy_input = true;
387 lex_error (lexer, _("expecting output value"));
393 /* Sets OUT as a numeric mapping output with the given VALUE. */
395 set_map_out_num (struct map_out *out, double value)
397 out->copy_input = false;
398 out->value.f = value;
402 /* Sets OUT as a string mapping output with the given VALUE. */
404 set_map_out_str (struct map_out *out, struct pool *pool,
405 const struct string *value)
407 const char *string = ds_data (value);
408 size_t length = ds_length (value);
410 out->copy_input = false;
411 out->value.c = pool_alloc_unaligned (pool, length);
412 memcpy (out->value.c, string, length);
416 /* Parses a set of target variables into TRNS->dst_vars and
419 parse_dst_vars (struct lexer *lexer, struct recode_trns *trns,
420 const struct dictionary *dict)
424 if (lex_match_id (lexer, "INTO"))
429 if (!parse_mixed_vars_pool (lexer, dict, trns->pool,
430 &trns->dst_names, &name_cnt,
434 if (name_cnt != trns->var_cnt)
436 msg (SE, _("%u variable(s) cannot be recoded into "
437 "%u variable(s). Specify the same number "
438 "of variables as source and target variables."),
439 (unsigned) trns->var_cnt, (unsigned) name_cnt);
443 trns->dst_vars = pool_nalloc (trns->pool,
444 trns->var_cnt, sizeof *trns->dst_vars);
445 for (i = 0; i < trns->var_cnt; i++)
448 v = trns->dst_vars[i] = dict_lookup_var (dict, trns->dst_names[i]);
449 if (v == NULL && trns->dst_type == ALPHA)
451 msg (SE, _("There is no variable named "
452 "%s. (All string variables specified "
453 "on INTO must already exist. Use the "
454 "STRING command to create a string "
463 trns->dst_vars = trns->src_vars;
464 if (trns->src_type != trns->dst_type)
466 msg (SE, _("INTO is required with %s input values "
467 "and %s output values."),
468 var_type_adj (trns->src_type),
469 var_type_adj (trns->dst_type));
474 for (i = 0; i < trns->var_cnt; i++)
476 struct variable *v = trns->dst_vars[i];
477 if (v != NULL && v->type != trns->dst_type)
479 msg (SE, _("Type mismatch. Cannot store %s data in "
481 trns->dst_type == ALPHA ? _("string") : _("numeric"),
482 v->type == ALPHA ? _("string") : _("numeric"),
491 /* Ensures that all the output values in TRNS are as wide as the
492 widest destination variable. */
494 enlarge_dst_widths (struct recode_trns *trns)
496 size_t max_dst_width;
500 for (i = 0; i < trns->var_cnt; i++)
502 struct variable *v = trns->dst_vars[i];
503 if (v->width > max_dst_width)
504 max_dst_width = v->width;
507 for (i = 0; i < trns->map_cnt; i++)
509 struct map_out *out = &trns->mappings[i].out;
510 if (!out->copy_input && out->width < max_dst_width)
512 char *s = pool_alloc_unaligned (trns->pool, max_dst_width + 1);
513 str_copy_rpad (s, max_dst_width + 1, out->value.c);
519 /* Creates destination variables that don't already exist. */
521 create_dst_vars (struct recode_trns *trns, struct dictionary *dict)
525 for (i = 0; i < trns->var_cnt; i++)
527 struct variable **var = &trns->dst_vars[i];
528 const char *name = trns->dst_names[i];
530 *var = dict_lookup_var (dict, name);
532 *var = dict_create_var_assert (dict, name, 0);
533 assert ((*var)->type == trns->dst_type);
537 /* Data transformation. */
539 /* Returns the output mapping in TRNS for an input of VALUE on
540 variable V, or a null pointer if there is no mapping. */
541 static const struct map_out *
542 find_src_numeric (struct recode_trns *trns, double value, struct variable *v)
546 for (m = trns->mappings; m < trns->mappings + trns->map_cnt; m++)
548 const struct map_in *in = &m->in;
549 const struct map_out *out = &m->out;
555 match = value == in->x.f;
558 match = mv_is_num_user_missing (&v->miss, value);
561 match = value >= in->x.f && value <= in->y.f;
577 /* Returns the output mapping in TRNS for an input of VALUE with
578 the given WIDTH, or a null pointer if there is no mapping. */
579 static const struct map_out *
580 find_src_string (struct recode_trns *trns, const char *value, int width)
584 for (m = trns->mappings; m < trns->mappings + trns->map_cnt; m++)
586 const struct map_in *in = &m->in;
587 struct map_out *out = &m->out;
593 match = !memcmp (value, in->x.c, width);
603 match = data_in (ss_buffer (value, width), FMT_F, 0, 0, &uv, 0);
619 /* Performs RECODE transformation. */
621 recode_trns_proc (void *trns_, struct ccase *c, casenumber case_idx UNUSED)
623 struct recode_trns *trns = trns_;
626 for (i = 0; i < trns->var_cnt; i++)
628 struct variable *src_var = trns->src_vars[i];
629 struct variable *dst_var = trns->dst_vars[i];
631 const union value *src_data = case_data (c, src_var->fv);
632 union value *dst_data = case_data_rw (c, dst_var->fv);
634 const struct map_out *out;
636 if (trns->src_type == NUMERIC)
637 out = find_src_numeric (trns, src_data->f, src_var);
639 out = find_src_string (trns, src_data->s, src_var->width);
641 if (trns->dst_type == NUMERIC)
644 dst_data->f = !out->copy_input ? out->value.f : src_data->f;
645 else if (trns->src_vars != trns->dst_vars)
646 dst_data->f = SYSMIS;
652 if (!out->copy_input)
653 memcpy (dst_data->s, out->value.c, dst_var->width);
654 else if (trns->src_vars != trns->dst_vars)
655 buf_copy_rpad (dst_data->s, dst_var->width,
656 src_data->s, src_var->width);
658 else if (trns->src_vars != trns->dst_vars)
659 memset (dst_data->s, ' ', dst_var->width);
663 return TRNS_CONTINUE;
666 /* Frees a RECODE transformation. */
668 recode_trns_free (void *trns_)
670 struct recode_trns *trns = trns_;
671 pool_destroy (trns->pool);