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
25 #include "bitvector.h"
26 #include "dictionary.h"
33 /* Parses a name as a variable within VS and returns the
34 variable's index if successful. On failure emits an error
35 message and returns a null pointer. */
37 parse_vs_variable_idx (const struct var_set *vs)
43 lex_error (_("expecting variable name"));
47 idx = var_set_lookup_var_idx (vs, tokid);
49 msg (SE, _("%s is not a variable name."), tokid);
55 /* Parses a name as a variable within VS and returns the variable
56 if successful. On failure emits an error message and returns
58 static struct variable *
59 parse_vs_variable (const struct var_set *vs)
61 int idx = parse_vs_variable_idx (vs);
62 return idx >= 0 ? var_set_get_var (vs, idx) : NULL;
65 /* Parses a variable name in dictionary D and returns the
66 variable if successful. On failure emits an error message and
67 returns a null pointer. */
69 parse_dict_variable (const struct dictionary *d)
71 struct var_set *vs = var_set_create_from_dict (d);
72 struct variable *var = parse_vs_variable (vs);
77 /* Parses a variable name in default_dict and returns the
78 variable if successful. On failure emits an error message and
79 returns a null pointer. */
83 return parse_dict_variable (default_dict);
86 /* Returns the dictionary class corresponding to a variable named
89 dict_class_from_id (const char *name)
91 assert (name != NULL);
104 /* Returns the name of dictionary class DICT_CLASS. */
106 dict_class_to_name (enum dict_class dict_class)
111 return _("ordinary");
122 /* Parses a set of variables from dictionary D given options
123 OPTS. Resulting list of variables stored in *VAR and the
124 number of variables into *CNT. Returns nonzero only if
127 parse_variables (const struct dictionary *d, struct variable ***var,
134 assert (var != NULL);
135 assert (cnt != NULL);
138 vs = var_set_create_from_dict (d);
139 success = parse_var_set_vars (vs, var, cnt, opts);
142 var_set_destroy (vs);
146 /* Parses a variable name from VS. If successful, sets *IDX to
147 the variable's index in VS, *CLASS to the variable's
148 dictionary class, and returns nonzero. Returns zero on
151 parse_var_idx_class (const struct var_set *vs, int *idx,
152 enum dict_class *class)
154 *idx = parse_vs_variable_idx (vs);
158 *class = dict_class_from_id (var_set_get_var (vs, *idx)->name);
162 /* Add the variable from VS with index IDX to the list of
163 variables V that has *NV elements and room for *MV.
164 Uses and updates INCLUDED to avoid duplicates if indicated by
165 PV_OPTS, which also affects what variables are allowed in
168 add_variable (struct variable ***v, int *nv, int *mv,
169 char *included, int pv_opts,
170 const struct var_set *vs, int idx)
172 struct variable *add = var_set_get_var (vs, idx);
174 if ((pv_opts & PV_NUMERIC) && add->type != NUMERIC)
175 msg (SW, _("%s is not a numeric variable. It will not be "
176 "included in the variable list."), add->name);
177 else if ((pv_opts & PV_STRING) && add->type != ALPHA)
178 msg (SE, _("%s is not a string variable. It will not be "
179 "included in the variable list."), add->name);
180 else if ((pv_opts & PV_NO_SCRATCH)
181 && dict_class_from_id (add->name) == DC_SCRATCH)
182 msg (SE, _("Scratch variables (such as %s) are not allowed "
183 "here."), add->name);
184 else if ((pv_opts & PV_SAME_TYPE) && *nv && add->type != (*v)[0]->type)
185 msg (SE, _("%s and %s are not the same type. All variables in "
186 "this variable list must be of the same type. %s "
187 "will be omitted from list."),
188 (*v)[0]->name, add->name, add->name);
189 else if ((pv_opts & PV_NO_DUPLICATE) && included[idx])
190 msg (SE, _("Variable %s appears twice in variable list."), add->name);
196 *v = xrealloc (*v, *mv * sizeof **v);
199 if ((pv_opts & PV_DUPLICATE) || !included[idx])
202 if (!(pv_opts & PV_DUPLICATE))
208 /* Adds the variables in VS with indexes FIRST_IDX through
209 LAST_IDX, inclusive, to the list of variables V that has *NV
210 elements and room for *MV. Uses and updates INCLUDED to avoid
211 duplicates if indicated by PV_OPTS, which also affects what
212 variables are allowed in appropriate ways. */
214 add_variables (struct variable ***v, int *nv, int *mv, char *included,
216 const struct var_set *vs, int first_idx, int last_idx,
217 enum dict_class class)
221 for (i = first_idx; i <= last_idx; i++)
222 if (dict_class_from_id (var_set_get_var (vs, i)->name) == class)
223 add_variable (v, nv, mv, included, pv_opts, vs, i);
226 /* Note that if parse_variables() returns 0, *v is free()'d.
227 Conversely, if parse_variables() returns non-zero, then *nv is
228 nonzero and *v is non-NULL. */
230 parse_var_set_vars (const struct var_set *vs,
231 struct variable ***v, int *nv,
241 /* At most one of PV_NUMERIC, PV_STRING, PV_SAME_TYPE may be
243 assert ((((pv_opts & PV_NUMERIC) != 0)
244 + ((pv_opts & PV_STRING) != 0)
245 + ((pv_opts & PV_SAME_TYPE) != 0)) <= 1);
247 /* PV_DUPLICATE and PV_NO_DUPLICATE are incompatible. */
248 assert (!(pv_opts & PV_DUPLICATE) || !(pv_opts & PV_NO_DUPLICATE));
250 if (!(pv_opts & PV_APPEND))
259 if (!(pv_opts & PV_DUPLICATE))
263 included = xcalloc (var_set_get_cnt (vs));
264 for (i = 0; i < *nv; i++)
265 included[(*v)[i]->index] = 1;
270 if (lex_match (T_ALL))
271 add_variables (v, nv, &mv, included, pv_opts,
272 vs, 0, var_set_get_cnt (vs) - 1, DC_ORDINARY);
277 enum dict_class class;
280 if (!parse_var_idx_class (vs, &first_idx, &class))
283 if (!lex_match (T_TO))
284 add_variable (v, nv, &mv, included, pv_opts,
289 enum dict_class last_class;
290 struct variable *first_var, *last_var;
292 if (!parse_var_idx_class (vs, &last_idx, &last_class))
295 first_var = var_set_get_var (vs, first_idx);
296 last_var = var_set_get_var (vs, last_idx);
298 if (last_idx < first_idx)
300 msg (SE, _("%s TO %s is not valid syntax since %s "
301 "precedes %s in the dictionary."),
302 first_var->name, last_var->name,
303 first_var->name, last_var->name);
306 if (class != last_class)
308 msg (SE, _("When using the TO keyword to specify several "
309 "variables, both variables must be from "
310 "the same variable dictionaries, of either "
311 "ordinary, scratch, or system variables. "
312 "%s is a %s variable, whereas %s is %s."),
313 first_var->name, dict_class_to_name (class),
314 last_var->name, dict_class_to_name (last_class));
318 add_variables (v, nv, &mv, included, pv_opts,
319 vs, first_idx, last_idx, class);
321 if (pv_opts & PV_SINGLE)
325 while (token == T_ID && var_set_lookup_var (vs, tokid) != NULL);
342 /* Extracts a numeric suffix from variable name S, copying it
343 into string R. Sets *D to the length of R and *N to its
346 extract_num (char *s, char *r, int *n, int *d)
350 /* Find first digit. */
351 cp = s + strlen (s) - 1;
352 while (isdigit ((unsigned char) *cp) && cp > s)
357 strncpy (r, s, cp - s);
360 /* Count initial zeros. */
369 while (isdigit ((unsigned char) *cp))
372 *n = (*n * 10) + (*cp - '0');
377 if (*n == 0 && *d == 0)
379 msg (SE, _("incorrect use of TO convention"));
385 /* Parses a list of variable names according to the DATA LIST version
386 of the TO convention. */
388 parse_DATA_LIST_vars (char ***names, int *nnames, int pv_opts)
398 assert (names != NULL);
399 assert (nnames != NULL);
400 assert ((pv_opts & ~(PV_APPEND | PV_SINGLE
401 | PV_NO_SCRATCH | PV_NO_DUPLICATE)) == 0);
402 /* FIXME: PV_NO_DUPLICATE is not implemented. */
404 if (pv_opts & PV_APPEND)
405 nvar = mvar = *nnames;
412 name1 = xmalloc (4 * (SHORT_NAME_LEN + 1));
413 name2 = &name1[1 * SHORT_NAME_LEN + 1];
414 root1 = &name1[2 * SHORT_NAME_LEN + 1];
415 root2 = &name1[3 * SHORT_NAME_LEN + 1];
420 lex_error ("expecting variable name");
423 if (dict_class_from_id (tokid) == DC_SCRATCH
424 && (pv_opts & PV_NO_SCRATCH))
426 msg (SE, _("Scratch variables not allowed here."));
429 strcpy (name1, tokid);
436 lex_error ("expecting variable name");
439 strcpy (name2, tokid);
442 if (!extract_num (name1, root1, &n1, &d1)
443 || !extract_num (name2, root2, &n2, &d2))
446 if (strcmp (root1, root2))
448 msg (SE, _("Prefixes don't match in use of TO convention."));
453 msg (SE, _("Bad bounds in use of TO convention."));
459 if (mvar < nvar + (n2 - n1 + 1))
461 mvar += ROUND_UP (n2 - n1 + 1, 16);
462 *names = xrealloc (*names, mvar * sizeof **names);
465 for (n = n1; n <= n2; n++)
467 (*names)[nvar] = xmalloc (SHORT_NAME_LEN + 1);
468 sprintf ((*names)[nvar], "%s%0*d", root1, d1, n);
477 *names = xrealloc (*names, mvar * sizeof **names);
479 (*names)[nvar++] = xstrdup (name1);
484 if (pv_opts & PV_SINGLE)
487 while (token == T_ID);
496 for (i = 0; i < nvar; i++)
505 /* Parses a list of variables where some of the variables may be
506 existing and the rest are to be created. Same args as
507 parse_DATA_LIST_vars(). */
509 parse_mixed_vars (char ***names, int *nnames, int pv_opts)
513 assert (names != NULL);
514 assert (nnames != NULL);
515 assert ((pv_opts & ~PV_APPEND) == 0);
517 if (!(pv_opts & PV_APPEND))
522 while (token == T_ID || token == T_ALL)
524 if (token == T_ALL || dict_lookup_var (default_dict, tokid) != NULL)
529 if (!parse_variables (default_dict, &v, &nv, PV_NONE))
531 *names = xrealloc (*names, (*nnames + nv) * sizeof **names);
532 for (i = 0; i < nv; i++)
533 (*names)[*nnames + i] = xstrdup (v[i]->name);
537 else if (!parse_DATA_LIST_vars (names, nnames, PV_APPEND))
543 for (i = 0; i < *nnames; i++)
544 free ((*names)[*nnames]);
551 /* A set of variables. */
554 size_t (*get_cnt) (const struct var_set *);
555 struct variable *(*get_var) (const struct var_set *, size_t idx);
556 int (*lookup_var_idx) (const struct var_set *, const char *);
557 void (*destroy) (struct var_set *);
561 /* Returns the number of variables in VS. */
563 var_set_get_cnt (const struct var_set *vs)
567 return vs->get_cnt (vs);
570 /* Return variable with index IDX in VS.
571 IDX must be less than the number of variables in VS. */
573 var_set_get_var (const struct var_set *vs, size_t idx)
576 assert (idx < var_set_get_cnt (vs));
578 return vs->get_var (vs, idx);
581 /* Returns the variable in VS named NAME, or a null pointer if VS
582 contains no variable with that name. */
584 var_set_lookup_var (const struct var_set *vs, const char *name)
586 int idx = var_set_lookup_var_idx (vs, name);
587 return idx >= 0 ? var_set_get_var (vs, idx) : NULL;
590 /* Returns the index in VS of the variable named NAME, or -1 if
591 VS contains no variable with that name. */
593 var_set_lookup_var_idx (const struct var_set *vs, const char *name)
596 assert (name != NULL);
597 assert (strlen (name) <= LONG_NAME_LEN );
599 return vs->lookup_var_idx (vs, name);
604 var_set_destroy (struct var_set *vs)
610 /* Returns the number of variables in VS. */
612 dict_var_set_get_cnt (const struct var_set *vs)
614 struct dictionary *d = vs->aux;
616 return dict_get_var_cnt (d);
619 /* Return variable with index IDX in VS.
620 IDX must be less than the number of variables in VS. */
621 static struct variable *
622 dict_var_set_get_var (const struct var_set *vs, size_t idx)
624 struct dictionary *d = vs->aux;
626 return dict_get_var (d, idx);
629 /* Returns the index of the variable in VS named NAME, or -1 if
630 VS contains no variable with that name. */
632 dict_var_set_lookup_var_idx (const struct var_set *vs, const char *name)
634 struct dictionary *d = vs->aux;
635 struct variable *v = dict_lookup_var (d, name);
636 return v != NULL ? v->index : -1;
641 dict_var_set_destroy (struct var_set *vs)
646 /* Returns a variable set based on D. */
648 var_set_create_from_dict (const struct dictionary *d)
650 struct var_set *vs = xmalloc (sizeof *vs);
651 vs->get_cnt = dict_var_set_get_cnt;
652 vs->get_var = dict_var_set_get_var;
653 vs->lookup_var_idx = dict_var_set_lookup_var_idx;
654 vs->destroy = dict_var_set_destroy;
655 vs->aux = (void *) d;
659 /* A variable set based on an array. */
662 struct variable *const *var;/* Array of variables. */
663 size_t var_cnt; /* Number of elements in var. */
664 struct hsh_table *name_tab; /* Hash from variable names to variables. */
665 struct hsh_table *longname_tab; /* Hash of short names indexed by long names */
668 /* Returns the number of variables in VS. */
670 array_var_set_get_cnt (const struct var_set *vs)
672 struct array_var_set *avs = vs->aux;
677 /* Return variable with index IDX in VS.
678 IDX must be less than the number of variables in VS. */
679 static struct variable *
680 array_var_set_get_var (const struct var_set *vs, size_t idx)
682 struct array_var_set *avs = vs->aux;
684 return (struct variable *) avs->var[idx];
687 /* Returns the index of the variable in VS named NAME, or -1 if
688 VS contains no variable with that name. */
690 array_var_set_lookup_var_idx (const struct var_set *vs, const char *name)
693 struct array_var_set *avs = vs->aux;
694 struct variable v, *vp, *const *vpp;
696 struct name_table_entry key;
699 struct name_table_entry *nte;
701 assert (avs->longname_tab);
704 nte = hsh_find (avs->longname_tab, &key);
709 short_name = nte->name;
711 strcpy (v.name, short_name);
713 vpp = hsh_find (avs->name_tab, &vp);
714 return vpp != NULL ? vpp - avs->var : -1;
719 array_var_set_destroy (struct var_set *vs)
721 struct array_var_set *avs = vs->aux;
723 hsh_destroy (avs->name_tab);
724 hsh_destroy (avs->longname_tab);
729 /* Returns a variable set based on the VAR_CNT variables in
732 var_set_create_from_array (struct variable *const *var, size_t var_cnt)
735 struct array_var_set *avs;
738 vs = xmalloc (sizeof *vs);
739 vs->get_cnt = array_var_set_get_cnt;
740 vs->get_var = array_var_set_get_var;
741 vs->lookup_var_idx = array_var_set_lookup_var_idx;
742 vs->destroy = array_var_set_destroy;
743 vs->aux = avs = xmalloc (sizeof *avs);
745 avs->var_cnt = var_cnt;
746 avs->name_tab = hsh_create (2 * var_cnt,
747 compare_var_ptr_names, hash_var_ptr_name,
750 avs->longname_tab = hsh_create (2 * var_cnt,
751 compare_long_names, hash_long_name,
752 (hsh_free_func *) free_nte,
755 for (i = 0; i < var_cnt; i++)
757 struct name_table_entry *nte ;
759 if (hsh_insert (avs->name_tab, &var[i]) != NULL)
761 var_set_destroy (vs);
765 nte = xmalloc (sizeof (*nte));
766 nte->name = strdup(var[i]->name);
767 nte->longname = strdup(var[i]->longname);
769 if (hsh_insert (avs->longname_tab, nte) != NULL)
771 var_set_destroy (vs);