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., 59 Temple Place - Suite 330, Boston, MA
28 /* Frequency tables. */
30 /* Frequency table entry. */
33 union value v; /* The value. */
34 double c; /* The number of occurrences of the value. */
37 /* Types of frequency tables. */
44 /* Entire frequency table. */
47 int mode; /* FRQM_GENERAL or FRQM_INTEGER. */
50 struct hsh_table *data; /* Undifferentiated data. */
53 double *vector; /* Frequencies proper. */
54 int min, max; /* The boundaries of the table. */
55 double out_of_range; /* Sum of weights of out-of-range values. */
56 double sysmis; /* Sum of weights of SYSMIS values. */
59 struct freq *valid; /* Valid freqs. */
60 int n_valid; /* Number of total freqs. */
62 struct freq *missing; /* Missing freqs. */
63 int n_missing; /* Number of missing freqs. */
66 double total_cases; /* Sum of weights of all cases. */
67 double valid_cases; /* Sum of weights of valid cases. */
70 /* Procedures' private per-variable data. */
72 /* Structure name suffixes for private data:
73 _proc: for a procedure (i.e., LIST -> list_proc).
74 _trns: for a transformation (i.e., COMPUTE -> compute_trns.
75 _pgm: for an input program (i.e., DATA LIST -> data_list_pgm). */
77 /* CROSSTABS private data. */
80 /* Integer mode only. */
81 int min; /* Minimum value. */
82 int max; /* Maximum value + 1. */
83 int count; /* max - min. */
87 /* FREQUENCIES private data. */
90 frq_mean = 0, frq_semean, frq_median, frq_mode, frq_stddev, frq_variance,
91 frq_kurt, frq_sekurt, frq_skew, frq_seskew, frq_range, frq_min, frq_max,
95 struct frequencies_proc
97 int used; /* 1=This variable already used. */
100 struct freq_tab tab; /* Frequencies table to use. */
103 int n_groups; /* Number of groups. */
104 double *groups; /* Groups. */
107 double stat[frq_n_stats];
110 /* LIST private data. */
113 int newline; /* Whether a new line begins here. */
114 int width; /* Field width. */
115 int vert; /* Whether to print the varname vertically. */
118 /* DESCRIPTIVES private data. Note that the DESCRIPTIVES procedure also
119 has a transformation, descriptives_trns. */
122 /* As these are used as bit indexes, there must be 32 or fewer.
123 Be very careful in adjusting these, see the structure below
124 and the table in descriptives.q. */
125 dsc_mean = 0, dsc_semean, dsc_stddev, dsc_variance, dsc_kurt,
126 dsc_sekurt, dsc_skew, dsc_seskew, dsc_range, dsc_min,
127 dsc_max, dsc_sum, dsc_n_stats
130 struct descriptives_proc
133 int dup; /* Finds duplicates in list of
135 char zname[10]; /* Name for z-score variable. */
138 double valid, miss; /* Valid, missing--general. */
140 /* Mean, moments about the mean. */
141 double X_bar, M2, M3, M4;
145 double stats[dsc_n_stats]; /* Everything glommed together. */
148 /* GET private data. */
151 int fv, nv; /* First, # of values. */
157 SRT_ASCEND, /* A, B, C, ..., X, Y, Z. */
158 SRT_DESCEND /* Z, Y, X, ..., C, B, A. */
161 /* SORT CASES private data. */
162 struct sort_cases_proc
164 int order; /* SRT_ASCEND or SRT_DESCEND. */
167 /* MEANS private data. */
170 double min, max; /* Range for integer mode. */
173 /* Different types of variables for MATRIX DATA procedure. Order is
174 important: these are used for sort keys. */
177 MXD_SPLIT, /* SPLIT FILE variables. */
178 MXD_ROWTYPE, /* ROWTYPE_. */
179 MXD_FACTOR, /* Factor variables. */
180 MXD_VARNAME, /* VARNAME_. */
181 MXD_CONTINUOUS, /* Continuous variables. */
186 /* MATRIX DATA private data. */
187 struct matrix_data_proc
189 int vartype; /* Variable type. */
190 int subtype; /* Subtype. */
193 /* MATCH FILES private data. */
194 struct match_files_proc
196 struct variable *master; /* Corresponding master file variable. */
200 /* Script variables. */
205 NUMERIC, /* A numeric variable. */
206 ALPHA /* A string variable. (STRING is pre-empted by lexer.h) */
209 /* Types of missing values. Order is significant, see
210 mis-val.c:parse_numeric(), sfm-read.c:sfm_read_dictionary()
211 sfm-write.c:sfm_write_dictionary(),
212 sysfile-info.c:cmd_sysfile_info(), mis-val.c:copy_missing_values(),
213 pfm-read.c:read_variables(), pfm-write.c:write_variables(),
214 apply-dict.c:cmd_apply_dictionary(), and more (?). */
217 MISSING_NONE, /* No user-missing values. */
218 MISSING_1, /* One user-missing value. */
219 MISSING_2, /* Two user-missing values. */
220 MISSING_3, /* Three user-missing values. */
221 MISSING_RANGE, /* [a,b]. */
222 MISSING_LOW, /* (-inf,a]. */
223 MISSING_HIGH, /* (a,+inf]. */
224 MISSING_RANGE_1, /* [a,b], c. */
225 MISSING_LOW_1, /* (-inf,a], b. */
226 MISSING_HIGH_1, /* (a,+inf), b. */
230 /* A variable's dictionary entry. */
233 char name[9]; /* As a string. */
234 int index; /* Index into its dictionary's var[]. */
235 int type; /* NUMERIC or ALPHA. */
237 /* Also important but parse_variables() doesn't need it. Still,
238 check before reordering. */
239 int width; /* Size of string variables in chars. */
240 int fv, nv; /* Index into `value's, number of values. */
241 unsigned init : 1; /* 1=VFM must init and possibly reinit. */
242 unsigned reinit : 1; /* Cases are: 1=reinitialized; 0=left. */
244 /* Missing values. */
245 int miss_type; /* One of the MISSING_* constants. */
246 union value missing[3]; /* User-missing value. */
248 /* Display formats. */
249 struct fmt_spec print; /* Default format for PRINT. */
250 struct fmt_spec write; /* Default format for WRITE. */
253 struct val_labs *val_labs;
254 char *label; /* Variable label. */
256 /* Per-procedure info. */
261 struct crosstab_proc crs;
262 struct descriptives_proc dsc;
263 struct frequencies_proc frq;
264 struct list_proc lst;
265 struct means_proc mns;
266 struct sort_cases_proc srt;
267 struct matrix_data_proc mxd;
268 struct match_files_proc mtf;
269 struct t_test_proc t_t;
274 int compare_variables (const void *, const void *, void *);
275 unsigned hash_variable (const void *, void *);
277 /* Classes of variables. */
280 DC_ORDINARY, /* Ordinary identifier. */
281 DC_SYSTEM, /* System variable. */
282 DC_SCRATCH /* Scratch variable. */
285 enum dict_class dict_class_from_id (const char *name);
286 const char *dict_class_to_name (enum dict_class dict_class);
288 /* Vector of variables. */
291 int idx; /* Index for dict_get_vector(). */
292 char name[9]; /* Name. */
293 struct variable **var; /* Vector of variables. */
294 int cnt; /* Number of variables. */
299 /* A single case. (This doesn't need to be a struct anymore, but it
300 remains so for hysterical raisins.) */
308 /* Complete dictionary state. */
311 struct dictionary *dict_create (void);
312 struct dictionary *dict_clone (const struct dictionary *);
313 void dict_clear (struct dictionary *);
314 void dict_destroy (struct dictionary *);
316 size_t dict_get_var_cnt (const struct dictionary *);
317 struct variable *dict_get_var (const struct dictionary *, size_t idx);
318 void dict_get_vars (const struct dictionary *,
319 struct variable ***vars, size_t *cnt,
320 unsigned exclude_classes);
322 struct variable *dict_create_var (struct dictionary *, const char *,
324 struct variable *dict_create_var_assert (struct dictionary *, const char *,
326 struct variable *dict_clone_var (struct dictionary *, const struct variable *,
328 void dict_rename_var (struct dictionary *, struct variable *, const char *);
330 struct variable *dict_lookup_var (const struct dictionary *, const char *);
331 struct variable *dict_lookup_var_assert (const struct dictionary *,
333 int dict_contains_var (const struct dictionary *, const struct variable *);
334 void dict_delete_var (struct dictionary *, struct variable *);
335 void dict_delete_vars (struct dictionary *,
336 struct variable *const *, size_t count);
337 void dict_reorder_vars (struct dictionary *,
338 struct variable *const *, size_t count);
339 int dict_rename_vars (struct dictionary *,
340 struct variable **, char **new_names,
341 size_t count, char **err_name);
343 struct variable *dict_get_weight (const struct dictionary *);
344 double dict_get_case_weight (const struct dictionary *, const struct ccase *);
345 void dict_set_weight (struct dictionary *, struct variable *);
347 struct variable *dict_get_filter (const struct dictionary *);
348 void dict_set_filter (struct dictionary *, struct variable *);
350 int dict_get_case_limit (const struct dictionary *);
351 void dict_set_case_limit (struct dictionary *, int);
353 int dict_get_next_value_idx (const struct dictionary *);
354 size_t dict_get_case_size (const struct dictionary *);
355 void dict_compact_values (struct dictionary *);
357 struct variable *const *dict_get_split_vars (const struct dictionary *);
358 size_t dict_get_split_cnt (const struct dictionary *);
359 void dict_set_split_vars (struct dictionary *,
360 struct variable *const *, size_t cnt);
362 const char *dict_get_label (const struct dictionary *);
363 void dict_set_label (struct dictionary *, const char *);
365 const char *dict_get_documents (const struct dictionary *);
366 void dict_set_documents (struct dictionary *, const char *);
368 int dict_create_vector (struct dictionary *,
370 struct variable **, size_t cnt);
371 const struct vector *dict_get_vector (const struct dictionary *,
373 size_t dict_get_vector_cnt (const struct dictionary *);
374 const struct vector *dict_lookup_vector (const struct dictionary *,
376 void dict_clear_vectors (struct dictionary *);
378 void discard_variables (void);
380 /* This is the active file dictionary. */
381 extern struct dictionary *default_dict;
383 /* Transformation state. */
385 /* Default file handle for DATA LIST, REREAD, REPEATING DATA
387 extern struct file_handle *default_handle;
389 /* PROCESS IF expression. */
390 extern struct expression *process_if_expr;
392 /* TEMPORARY support. */
394 /* 1=TEMPORARY has been executed at some point. */
395 extern int temporary;
397 /* If temporary!=0, the saved dictionary. */
398 extern struct dictionary *temp_dict;
400 /* If temporary!=0, index into t_trns[] (declared far below) that
401 gives the point at which data should be written out. -1 means that
402 the data shouldn't be changed since all transformations are
404 extern int temp_trns;
406 /* If FILTER is active, whether it was executed before or after
408 extern int FILTER_before_TEMPORARY;
410 void cancel_temporary (void);
414 void dump_split_vars (const struct ccase *);
416 int is_num_user_missing (double, const struct variable *);
417 int is_str_user_missing (const unsigned char[], const struct variable *);
418 int is_missing (const union value *, const struct variable *);
419 int is_system_missing (const union value *, const struct variable *);
420 int is_user_missing (const union value *, const struct variable *);
421 void copy_missing_values (struct variable *dest, const struct variable *src);
423 /* Transformations. */
425 /* Header for all transformations. */
428 /* Index into t_trns[]. */
431 /* Transformation proc. */
432 int (*proc) (struct trns_header *, struct ccase *);
434 /* Garbage collector proc. */
435 void (*free) (struct trns_header *);
438 /* Array of transformations */
439 extern struct trns_header **t_trns;
441 /* Number of transformations, maximum number in array currently. */
442 extern int n_trns, m_trns;
444 /* Index of first transformation that is really a transformation. Any
445 transformations before this belong to INPUT PROGRAM. */
448 void add_transformation (struct trns_header *trns);
449 void cancel_transformations (void);
453 struct var_set *var_set_create_from_dict (struct dictionary *d);
454 struct var_set *var_set_create_from_array (struct variable **var, size_t);
456 size_t var_set_get_cnt (struct var_set *vs);
457 struct variable *var_set_get_var (struct var_set *vs, size_t idx);
458 struct variable *var_set_lookup_var (struct var_set *vs, const char *name);
459 void var_set_destroy (struct var_set *vs);
462 /* Variable parsers. */
466 PV_NONE = 0, /* No options. */
467 PV_SINGLE = 0001, /* Restrict to a single name or TO use. */
468 PV_DUPLICATE = 0002, /* Don't merge duplicates. */
469 PV_APPEND = 0004, /* Append to existing list. */
470 PV_NO_DUPLICATE = 0010, /* Error on duplicates. */
471 PV_NUMERIC = 0020, /* Vars must be numeric. */
472 PV_STRING = 0040, /* Vars must be string. */
473 PV_SAME_TYPE = 00100, /* All vars must be the same type. */
474 PV_NO_SCRATCH = 00200, /* Disallow scratch variables. */
477 struct variable *parse_variable (void);
478 struct variable *parse_dict_variable (struct dictionary *);
479 int parse_variables (struct dictionary *, struct variable ***, int *,
481 int parse_var_set_vars (struct var_set *, struct variable ***, int *,
483 int parse_DATA_LIST_vars (char ***names, int *cnt, int opts);
484 int parse_mixed_vars (char ***names, int *cnt, int opts);