X-Git-Url: https://pintos-os.org/cgi-bin/gitweb.cgi?a=blobdiff_plain;f=src%2Flanguage%2Fstats%2Fautorecode.c;h=33cab462c4d6ab226480722fe31cae218bdda6ff;hb=15aae7ade3a0835939d28d78f20ed1eb85500d86;hp=05016548e37b4d98cb6cedb6e3416f891cf763e3;hpb=d500232dbfbf76be9e525e3e5cdf59505b0cdf32;p=pspp diff --git a/src/language/stats/autorecode.c b/src/language/stats/autorecode.c index 05016548e3..33cab462c4 100644 --- a/src/language/stats/autorecode.c +++ b/src/language/stats/autorecode.c @@ -20,24 +20,25 @@ #include "data/case.h" #include "data/casereader.h" +#include "data/dataset.h" #include "data/dictionary.h" -#include "data/procedure.h" #include "data/transformations.h" #include "data/variable.h" #include "language/command.h" #include "language/lexer/lexer.h" #include "language/lexer/variable-parser.h" #include "libpspp/array.h" -#include "libpspp/i18n.h" #include "libpspp/compiler.h" #include "libpspp/hash-functions.h" #include "libpspp/hmap.h" +#include "libpspp/i18n.h" #include "libpspp/message.h" #include "libpspp/pool.h" #include "libpspp/str.h" #include "gl/xalloc.h" #include "gl/vasnprintf.h" +#include "gl/mbiter.h" #include "gettext.h" #define _(msgid) gettext (msgid) @@ -49,6 +50,8 @@ struct arc_item { struct hmap_node hmap_node; /* Element in "struct arc_spec" hash table. */ union value from; /* Original value. */ + int width; /* Width of the original value */ + double to; /* Recoded value. */ }; @@ -75,23 +78,48 @@ struct autorecode_pgm /* Hash table of "struct arc_item"s. */ struct hmap *global_items; + + bool blank_valid; + const struct dictionary *dict; }; static trns_proc_func autorecode_trns_proc; static trns_free_func autorecode_trns_free; -static int compare_arc_items (const void *, const void *, const void *width); +static int compare_arc_items (const void *, const void *, const void *aux); static void arc_free (struct autorecode_pgm *); -static struct arc_item *find_arc_item (struct arc_spec *, const union value *, +static struct arc_item *find_arc_item (const struct arc_spec *, const union value *, size_t hash); +static bool +value_is_blank (const union value *val, int width, const struct dictionary *dict) +{ + mbi_iterator_t iter; + const char *str = CHAR_CAST_BUG (const char*, value_str (val, width)); + char *text = recode_string (UTF8, dict_get_encoding (dict), str, width); + + for (mbi_init (iter, text, width); mbi_avail (iter); mbi_advance (iter)) + { + mbchar_t c = mbi_cur (iter); + + if ( ! mb_isblank (c)) + { + free (text); + return false; + } + } + + free (text); + return true; +} + /* Performs the AUTORECODE procedure. */ int cmd_autorecode (struct lexer *lexer, struct dataset *ds) { struct autorecode_pgm *arc = NULL; - struct dictionary *dict = dataset_dict (ds); + struct dictionary *dict = dataset_dict (ds); const struct variable **src_vars = NULL; char **dst_names = NULL; size_t n_srcs = 0; @@ -108,17 +136,20 @@ cmd_autorecode (struct lexer *lexer, struct dataset *ds) /* Create procedure. */ arc = xzalloc (sizeof *arc); + arc->blank_valid = true; + arc->dict = dataset_dict (ds); /* Parse variable lists. */ lex_match_id (lexer, "VARIABLES"); - lex_match (lexer, '='); - if (!parse_variables_const (lexer, dict, &src_vars, &n_srcs, + lex_match (lexer, T_EQUALS); + if (!parse_variables_const (lexer, arc->dict, &src_vars, &n_srcs, PV_NO_DUPLICATE)) goto error; if (!lex_force_match_id (lexer, "INTO")) goto error; - lex_match (lexer, '='); - if (!parse_DATA_LIST_vars (lexer, &dst_names, &n_dsts, PV_NO_DUPLICATE)) + lex_match (lexer, T_EQUALS); + if (!parse_DATA_LIST_vars (lexer, arc->dict, &dst_names, &n_dsts, + PV_NO_DUPLICATE)) goto error; if (n_dsts != n_srcs) { @@ -132,7 +163,7 @@ cmd_autorecode (struct lexer *lexer, struct dataset *ds) { const char *name = dst_names[i]; - if (dict_lookup_var (dict, name) != NULL) + if (dict_lookup_var (arc->dict, name) != NULL) { msg (SE, _("Target variable %s duplicates existing variable %s."), name, name); @@ -141,7 +172,7 @@ cmd_autorecode (struct lexer *lexer, struct dataset *ds) } /* Parse options. */ - while (lex_match (lexer, '/')) + while (lex_match (lexer, T_SLASH)) { if (lex_match_id (lexer, "DESCENDING")) direction = DESCENDING; @@ -152,9 +183,25 @@ cmd_autorecode (struct lexer *lexer, struct dataset *ds) arc->global_items = xmalloc (sizeof (*arc->global_items)); hmap_init (arc->global_items); } + else if (lex_match_id (lexer, "BLANK")) + { + lex_match (lexer, T_EQUALS); + if (lex_match_id (lexer, "VALID")) + { + arc->blank_valid = true; + } + else if (lex_match_id (lexer, "MISSING")) + { + arc->blank_valid = false; + } + else + goto error; + } + else + goto error; } - if (lex_token (lexer) != '.') + if (lex_token (lexer) != T_ENDCMD) { lex_error (lexer, _("expecting end of command")); goto error; @@ -163,6 +210,7 @@ cmd_autorecode (struct lexer *lexer, struct dataset *ds) arc->specs = xmalloc (n_dsts * sizeof *arc->specs); arc->n_specs = n_dsts; + for (i = 0; i < n_dsts; i++) { struct arc_spec *spec = &arc->specs[i]; @@ -192,9 +240,13 @@ cmd_autorecode (struct lexer *lexer, struct dataset *ds) struct arc_item *item; item = find_arc_item (spec, value, hash); - if (item == NULL) + if ( (item == NULL) + && + ( arc->blank_valid || var_is_numeric (spec->src) || ! value_is_blank (value, width, arc->dict)) + ) { item = xmalloc (sizeof *item); + item->width = width; value_clone (&item->from, value, width); hmap_insert (spec->items, &item->hmap_node, hash); } @@ -209,7 +261,6 @@ cmd_autorecode (struct lexer *lexer, struct dataset *ds) struct arc_item **items; struct arc_item *item; size_t n_items; - int src_width; size_t j; /* Create destination variable. */ @@ -221,12 +272,11 @@ cmd_autorecode (struct lexer *lexer, struct dataset *ds) j = 0; HMAP_FOR_EACH (item, struct arc_item, hmap_node, spec->items) items[j++] = item; - if (!arc->global_items) - assert (j == n_items); + + assert (j == n_items); /* Sort array by value. */ - src_width = var_get_width (spec->src); - sort (items, n_items, sizeof *items, compare_arc_items, &src_width); + sort (items, n_items, sizeof *items, compare_arc_items, NULL); /* Assign recoded values in sorted order. */ for (j = 0; j < n_items; j++) @@ -235,7 +285,7 @@ cmd_autorecode (struct lexer *lexer, struct dataset *ds) size_t len; char *recoded_value = NULL; char *c; - + const int src_width = items[j]->width; union value to_val; value_init (&to_val, 0); @@ -245,12 +295,11 @@ cmd_autorecode (struct lexer *lexer, struct dataset *ds) /* Add value labels to the destination variable which indicate the source value from whence the new value comes. */ - if (src_width > 0) { - const char *str = (const char *) value_str (from, src_width); + const char *str = CHAR_CAST_BUG (const char*, value_str (from, src_width)); - recoded_value = recode_string (UTF8, dict_get_encoding (dict), str, src_width); + recoded_value = recode_string (UTF8, dict_get_encoding (arc->dict), str, src_width); } else recoded_value = asnprintf (NULL, &len, "%g", from->f); @@ -294,13 +343,12 @@ arc_free (struct autorecode_pgm *arc) for (i = 0; i < arc->n_specs; i++) { struct arc_spec *spec = &arc->specs[i]; - int width = var_get_width (spec->src); struct arc_item *item, *next; HMAP_FOR_EACH_SAFE (item, next, struct arc_item, hmap_node, spec->items) { - value_destroy (&item->from, width); + value_destroy (&item->from, item->width); hmap_delete (spec->items, &item->hmap_node); free (item); } @@ -329,7 +377,7 @@ arc_free (struct autorecode_pgm *arc) } static struct arc_item * -find_arc_item (struct arc_spec *spec, const union value *value, +find_arc_item (const struct arc_spec *spec, const union value *value, size_t hash) { struct arc_item *item; @@ -341,13 +389,24 @@ find_arc_item (struct arc_spec *spec, const union value *value, } static int -compare_arc_items (const void *a_, const void *b_, const void *width_) +compare_arc_items (const void *a_, const void *b_, const void *aux UNUSED) { const struct arc_item *const *a = a_; const struct arc_item *const *b = b_; - const int *width = width_; + int width_a = (*a)->width; + int width_b = (*b)->width; + + if ( width_a == width_b) + return value_compare_3way (&(*a)->from, &(*b)->from, width_a); - return value_compare_3way (&(*a)->from, &(*b)->from, *width); + if ( width_a == 0 && width_b != 0) + return -1; + + if ( width_b == 0 && width_a != 0) + return +1; + + return buf_compare_rpad (CHAR_CAST_BUG (const char *, value_str (&(*a)->from, width_a)), width_a, + CHAR_CAST_BUG (const char *, value_str (&(*b)->from, width_b)), width_b); } static int @@ -360,12 +419,11 @@ autorecode_trns_proc (void *arc_, struct ccase **c, *c = case_unshare (*c); for (i = 0; i < arc->n_specs; i++) { - struct arc_spec *spec = &arc->specs[i]; + const struct arc_spec *spec = &arc->specs[i]; int width = var_get_width (spec->src); const union value *value = case_data (*c, spec->src); - struct arc_item *item; + const struct arc_item *item = find_arc_item (spec, value, value_hash (value, width, 0)); - item = find_arc_item (spec, value, value_hash (value, width, 0)); case_data_rw (*c, spec->dst)->f = item ? item->to : SYSMIS; }