1 /* PSPP - a program for statistical analysis.
2 Copyright (C) 1997-9, 2000, 2009, 2010, 2011 Free Software Foundation, Inc.
4 This program is free software: you can redistribute it and/or modify
5 it under the terms of the GNU General Public License as published by
6 the Free Software Foundation, either version 3 of the License, or
7 (at your option) any later version.
9 This program is distributed in the hope that it will be useful,
10 but WITHOUT ANY WARRANTY; without even the implied warranty of
11 MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
12 GNU General Public License for more details.
14 You should have received a copy of the GNU General Public License
15 along with this program. If not, see <http://www.gnu.org/licenses/>. */
23 #include "data/case.h"
24 #include "data/data-in.h"
25 #include "data/dictionary.h"
26 #include "data/format.h"
27 #include "data/procedure.h"
28 #include "data/transformations.h"
29 #include "data/variable.h"
30 #include "language/command.h"
31 #include "language/lexer/lexer.h"
32 #include "language/lexer/value-parser.h"
33 #include "language/lexer/variable-parser.h"
34 #include "libpspp/assertion.h"
35 #include "libpspp/cast.h"
36 #include "libpspp/compiler.h"
37 #include "libpspp/i18n.h"
38 #include "libpspp/message.h"
39 #include "libpspp/pool.h"
40 #include "libpspp/str.h"
42 #include "gl/xalloc.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 /* Describes input values to be mapped. */
63 enum map_in_type type; /* One of MAP_*. */
64 union value x, y; /* Source values. */
67 /* Describes the value used as output from a mapping. */
70 bool copy_input; /* If true, copy input to output. */
71 union value value; /* If copy_input false, recoded value. */
72 int width; /* If copy_input false, output value width. */
75 /* Describes how to recode a single value or range of values into a
79 struct map_in in; /* Input values. */
80 struct map_out out; /* Output value. */
83 /* RECODE transformation. */
90 /* Variable types, for convenience. */
91 enum val_type src_type; /* src_vars[*] type. */
92 enum val_type dst_type; /* dst_vars[*] type. */
95 const struct variable **src_vars; /* Source variables. */
96 const struct variable **dst_vars; /* Destination variables. */
97 const struct dictionary *dst_dict; /* Dictionary of dst_vars */
98 char **dst_names; /* Name of dest variables, if they're new. */
99 size_t var_cnt; /* Number of variables. */
102 struct mapping *mappings; /* Value mappings. */
103 size_t map_cnt; /* Number of mappings. */
104 int max_src_width; /* Maximum width of src_vars[*]. */
105 int max_dst_width; /* Maximum width of any map_out in mappings. */
108 static bool parse_src_vars (struct lexer *, struct recode_trns *, const struct dictionary *dict);
109 static bool parse_mappings (struct lexer *, struct recode_trns *);
110 static bool parse_dst_vars (struct lexer *, struct recode_trns *, const struct dictionary *dict);
112 static void add_mapping (struct recode_trns *,
113 size_t *map_allocated, const struct map_in *);
115 static bool parse_map_in (struct lexer *lexer, struct map_in *, struct pool *,
116 enum val_type src_type, size_t max_src_width);
117 static void set_map_in_generic (struct map_in *, enum map_in_type);
118 static void set_map_in_num (struct map_in *, enum map_in_type, double, double);
119 static void set_map_in_str (struct map_in *, struct pool *,
120 struct substring, size_t width);
122 static bool parse_map_out (struct lexer *lexer, struct pool *, struct map_out *);
123 static void set_map_out_num (struct map_out *, double);
124 static void set_map_out_str (struct map_out *, struct pool *,
127 static void enlarge_dst_widths (struct recode_trns *);
128 static void create_dst_vars (struct recode_trns *, struct dictionary *);
130 static trns_proc_func recode_trns_proc;
131 static trns_free_func recode_trns_free;
135 /* Parses the RECODE transformation. */
137 cmd_recode (struct lexer *lexer, struct dataset *ds)
141 struct recode_trns *trns
142 = pool_create_container (struct recode_trns, pool);
144 /* Parse source variable names,
145 then input to output mappings,
146 then destintation variable names. */
147 if (!parse_src_vars (lexer, trns, dataset_dict (ds) )
148 || !parse_mappings (lexer, trns)
149 || !parse_dst_vars (lexer, trns, dataset_dict (ds)))
151 recode_trns_free (trns);
155 /* Ensure that all the output strings are at least as wide
156 as the widest destination variable. */
157 if (trns->dst_type == VAL_STRING)
158 enlarge_dst_widths (trns);
160 /* Create destination variables, if needed.
161 This must be the final step; otherwise we'd have to
162 delete destination variables on failure. */
163 trns->dst_dict = dataset_dict (ds);
164 if (trns->src_vars != trns->dst_vars)
165 create_dst_vars (trns, dataset_dict (ds));
168 add_transformation (ds,
169 recode_trns_proc, recode_trns_free, trns);
171 while (lex_match (lexer, T_SLASH));
173 return lex_end_of_command (lexer);
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 lexer *lexer,
181 struct recode_trns *trns, const struct dictionary *dict)
183 if (!parse_variables_const (lexer, dict, &trns->src_vars, &trns->var_cnt,
186 pool_register (trns->pool, free, trns->src_vars);
187 trns->src_type = var_get_type (trns->src_vars[0]);
191 /* Parses a set of mappings, which take the form (input=output),
192 into TRNS->mappings and TRNS->map_cnt. Sets TRNS->dst_type.
193 Returns true if successful, false on parse error. */
195 parse_mappings (struct lexer *lexer, struct recode_trns *trns)
197 size_t map_allocated;
201 /* Find length of longest source variable. */
202 trns->max_src_width = var_get_width (trns->src_vars[0]);
203 for (i = 1; i < trns->var_cnt; i++)
205 size_t var_width = var_get_width (trns->src_vars[i]);
206 if (var_width > trns->max_src_width)
207 trns->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 (lexer, T_LPAREN))
219 enum val_type dst_type;
221 if (!lex_match_id (lexer, "CONVERT"))
224 size_t first_map_idx;
227 first_map_idx = trns->map_cnt;
229 /* Parse source specifications. */
234 if (!parse_map_in (lexer, &in, trns->pool,
235 trns->src_type, trns->max_src_width))
237 add_mapping (trns, &map_allocated, &in);
238 lex_match (lexer, T_COMMA);
240 while (!lex_match (lexer, T_EQUALS));
242 if (!parse_map_out (lexer, trns->pool, &out))
246 dst_type = trns->src_type;
248 dst_type = val_type_from_width (out.width);
249 if (have_dst_type && dst_type != trns->dst_type)
251 msg (SE, _("Inconsistent target variable types. "
253 "must be all numeric or all string."));
257 for (i = first_map_idx; i < trns->map_cnt; i++)
258 trns->mappings[i].out = out;
262 /* Parse CONVERT as a special case. */
264 set_map_in_generic (&in, MAP_CONVERT);
265 add_mapping (trns, &map_allocated, &in);
266 set_map_out_num (&trns->mappings[trns->map_cnt - 1].out, 0.0);
268 dst_type = VAL_NUMERIC;
269 if (trns->src_type != VAL_STRING
270 || (have_dst_type && trns->dst_type != VAL_NUMERIC))
272 msg (SE, _("CONVERT requires string input values and "
273 "numeric output values."));
277 trns->dst_type = dst_type;
278 have_dst_type = true;
280 if (!lex_force_match (lexer, T_RPAREN))
283 while (lex_match (lexer, T_LPAREN));
288 /* Parses a mapping input value into IN, allocating memory from
289 POOL. The source value type must be provided as SRC_TYPE and,
290 if string, the maximum width of a string source variable must
291 be provided in MAX_SRC_WIDTH. Returns true if successful,
292 false on parse error. */
294 parse_map_in (struct lexer *lexer, struct map_in *in, struct pool *pool,
295 enum val_type src_type, size_t max_src_width)
298 if (lex_match_id (lexer, "ELSE"))
299 set_map_in_generic (in, MAP_ELSE);
300 else if (src_type == VAL_NUMERIC)
302 if (lex_match_id (lexer, "MISSING"))
303 set_map_in_generic (in, MAP_MISSING);
304 else if (lex_match_id (lexer, "SYSMIS"))
305 set_map_in_generic (in, MAP_SYSMIS);
309 if (!parse_num_range (lexer, &x, &y, NULL))
311 set_map_in_num (in, x == y ? MAP_SINGLE : MAP_RANGE, x, y);
316 if (lex_match_id (lexer, "MISSING"))
317 set_map_in_generic (in, MAP_MISSING);
318 else if (!lex_force_string (lexer))
322 set_map_in_str (in, pool, lex_tokss (lexer), max_src_width);
324 if (lex_token (lexer) == T_ID
325 && lex_id_match (ss_cstr ("THRU"), lex_tokss (lexer)))
327 msg (SE, _("THRU is not allowed with string variables."));
336 /* Adds IN to the list of mappings in TRNS.
337 MAP_ALLOCATED is the current number of allocated mappings,
338 which is updated as needed. */
340 add_mapping (struct recode_trns *trns,
341 size_t *map_allocated, const struct map_in *in)
344 if (trns->map_cnt >= *map_allocated)
345 trns->mappings = pool_2nrealloc (trns->pool, trns->mappings,
347 sizeof *trns->mappings);
348 m = &trns->mappings[trns->map_cnt++];
352 /* Sets IN as a mapping of the given TYPE. */
354 set_map_in_generic (struct map_in *in, enum map_in_type type)
359 /* Sets IN as a numeric mapping of the given TYPE,
360 with X and Y as the two numeric values. */
362 set_map_in_num (struct map_in *in, enum map_in_type type, double x, double y)
369 /* Sets IN as a string mapping, with STRING as the string,
370 allocated from POOL. The string is padded with spaces on the
371 right to WIDTH characters long. */
373 set_map_in_str (struct map_in *in, struct pool *pool,
374 struct substring string, size_t width)
376 in->type = MAP_SINGLE;
377 value_init_pool (pool, &in->x, width);
378 value_copy_buf_rpad (&in->x, width,
379 CHAR_CAST_BUG (uint8_t *, ss_data (string)),
380 ss_length (string), ' ');
383 /* Parses a mapping output value into OUT, allocating memory from
384 POOL. Returns true if successful, false on parse error. */
386 parse_map_out (struct lexer *lexer, struct pool *pool, struct map_out *out)
388 if (lex_is_number (lexer))
390 set_map_out_num (out, lex_number (lexer));
393 else if (lex_match_id (lexer, "SYSMIS"))
394 set_map_out_num (out, SYSMIS);
395 else if (lex_is_string (lexer))
397 set_map_out_str (out, pool, lex_tokss (lexer));
400 else if (lex_match_id (lexer, "COPY"))
402 out->copy_input = true;
407 lex_error (lexer, _("expecting output value"));
413 /* Sets OUT as a numeric mapping output with the given VALUE. */
415 set_map_out_num (struct map_out *out, double value)
417 out->copy_input = false;
418 out->value.f = value;
422 /* Sets OUT as a string mapping output with the given VALUE. */
424 set_map_out_str (struct map_out *out, struct pool *pool,
425 const struct substring value)
427 const char *string = ss_data (value);
428 size_t length = ss_length (value);
432 /* A length of 0 will yield a numeric value, which is not
438 out->copy_input = false;
439 value_init_pool (pool, &out->value, length);
440 memcpy (value_str_rw (&out->value, length), string, length);
444 /* Parses a set of target variables into TRNS->dst_vars and
447 parse_dst_vars (struct lexer *lexer, struct recode_trns *trns,
448 const struct dictionary *dict)
452 if (lex_match_id (lexer, "INTO"))
457 if (!parse_mixed_vars_pool (lexer, dict, trns->pool,
458 &trns->dst_names, &name_cnt,
462 if (name_cnt != trns->var_cnt)
464 msg (SE, _("%zu variable(s) cannot be recoded into "
465 "%zu variable(s). Specify the same number "
466 "of variables as source and target variables."),
467 trns->var_cnt, name_cnt);
471 trns->dst_vars = pool_nalloc (trns->pool,
472 trns->var_cnt, sizeof *trns->dst_vars);
473 for (i = 0; i < trns->var_cnt; i++)
475 const struct variable *v;
476 v = trns->dst_vars[i] = dict_lookup_var (dict, trns->dst_names[i]);
477 if (v == NULL && trns->dst_type == VAL_STRING)
479 msg (SE, _("There is no variable named "
480 "%s. (All string variables specified "
481 "on INTO must already exist. Use the "
482 "STRING command to create a string "
492 trns->dst_vars = trns->src_vars;
493 if (trns->src_type != trns->dst_type)
495 msg (SE, _("INTO is required with %s input values "
496 "and %s output values."),
497 trns->src_type == VAL_NUMERIC ? _("numeric") : _("string"),
498 trns->dst_type == VAL_NUMERIC ? _("numeric") : _("string"));
503 for (i = 0; i < trns->var_cnt; i++)
505 const struct variable *v = trns->dst_vars[i];
506 if (v != NULL && var_get_type (v) != trns->dst_type)
508 msg (SE, _("Type mismatch. Cannot store %s data in "
510 trns->dst_type == VAL_STRING ? _("string") : _("numeric"),
511 var_is_alpha (v) ? _("string") : _("numeric"),
520 /* Ensures that all the output values in TRNS are as wide as the
521 widest destination variable. */
523 enlarge_dst_widths (struct recode_trns *trns)
527 trns->max_dst_width = 0;
528 for (i = 0; i < trns->var_cnt; i++)
530 const struct variable *v = trns->dst_vars[i];
531 if (var_get_width (v) > trns->max_dst_width)
532 trns->max_dst_width = var_get_width (v);
535 for (i = 0; i < trns->map_cnt; i++)
537 struct map_out *out = &trns->mappings[i].out;
538 if (!out->copy_input)
539 value_resize_pool (trns->pool, &out->value,
540 out->width, trns->max_dst_width);
544 /* Creates destination variables that don't already exist. */
546 create_dst_vars (struct recode_trns *trns, struct dictionary *dict)
550 for (i = 0; i < trns->var_cnt; i++)
552 const struct variable **var = &trns->dst_vars[i];
553 const char *name = trns->dst_names[i];
555 *var = dict_lookup_var (dict, name);
557 *var = dict_create_var_assert (dict, name, 0);
558 assert (var_get_type (*var) == trns->dst_type);
562 /* Data transformation. */
564 /* Returns the output mapping in TRNS for an input of VALUE on
565 variable V, or a null pointer if there is no mapping. */
566 static const struct map_out *
567 find_src_numeric (struct recode_trns *trns, double value, const struct variable *v)
571 for (m = trns->mappings; m < trns->mappings + trns->map_cnt; m++)
573 const struct map_in *in = &m->in;
574 const struct map_out *out = &m->out;
580 match = value == in->x.f;
583 match = var_is_num_missing (v, value, MV_ANY);
586 match = value >= in->x.f && value <= in->y.f;
589 match = value == SYSMIS;
605 /* Returns the output mapping in TRNS for an input of VALUE with
606 the given WIDTH, or a null pointer if there is no mapping. */
607 static const struct map_out *
608 find_src_string (struct recode_trns *trns, const uint8_t *value,
609 const struct variable *src_var)
611 const char *encoding = dict_get_encoding (trns->dst_dict);
612 int width = var_get_width (src_var);
615 for (m = trns->mappings; m < trns->mappings + trns->map_cnt; m++)
617 const struct map_in *in = &m->in;
618 struct map_out *out = &m->out;
624 match = !memcmp (value, value_str (&in->x, trns->max_src_width),
635 error = data_in (ss_buffer (CHAR_CAST_BUG (char *, value), width),
636 C_ENCODING, FMT_F, &uv, 0, encoding);
637 match = error == NULL;
644 match = var_is_str_missing (src_var, value, MV_ANY);
657 /* Performs RECODE transformation. */
659 recode_trns_proc (void *trns_, struct ccase **c, casenumber case_idx UNUSED)
661 struct recode_trns *trns = trns_;
664 *c = case_unshare (*c);
665 for (i = 0; i < trns->var_cnt; i++)
667 const struct variable *src_var = trns->src_vars[i];
668 const struct variable *dst_var = trns->dst_vars[i];
669 const struct map_out *out;
671 if (trns->src_type == VAL_NUMERIC)
672 out = find_src_numeric (trns, case_num (*c, src_var), src_var);
674 out = find_src_string (trns, case_str (*c, src_var), src_var);
676 if (trns->dst_type == VAL_NUMERIC)
678 double *dst = &case_data_rw (*c, dst_var)->f;
680 *dst = !out->copy_input ? out->value.f : case_num (*c, src_var);
681 else if (trns->src_vars != trns->dst_vars)
686 char *dst = CHAR_CAST_BUG (char *, case_str_rw (*c, dst_var));
689 if (!out->copy_input)
690 memcpy (dst, value_str (&out->value, trns->max_dst_width),
691 var_get_width (dst_var));
692 else if (trns->src_vars != trns->dst_vars)
694 union value *dst_data = case_data_rw (*c, dst_var);
695 const union value *src_data = case_data (*c, src_var);
696 value_copy_rpad (dst_data, var_get_width (dst_var),
697 src_data, var_get_width (src_var), ' ');
700 else if (trns->src_vars != trns->dst_vars)
701 memset (dst, ' ', var_get_width (dst_var));
705 return TRNS_CONTINUE;
708 /* Frees a RECODE transformation. */
710 recode_trns_free (void *trns_)
712 struct recode_trns *trns = trns_;
713 pool_destroy (trns->pool);