X-Git-Url: https://pintos-os.org/cgi-bin/gitweb.cgi?a=blobdiff_plain;f=src%2Flanguage%2Fstats%2Foneway.q;h=78300723ab948525364cf41dad6c6333d919518b;hb=92c09e564002d356d20fc1e2e131027ef89f6748;hp=674cf710f085f4cbbeb2b81f07537921af86e6d7;hpb=18f6e8958244f938e9e9a03a4230cacf0d22a470;p=pspp-builds.git diff --git a/src/language/stats/oneway.q b/src/language/stats/oneway.q index 674cf710..78300723 100644 --- a/src/language/stats/oneway.q +++ b/src/language/stats/oneway.q @@ -1,7 +1,6 @@ /* PSPP - One way ANOVA. -*-c-*- -Copyright (C) 1997-9, 2000 Free Software Foundation, Inc. -Author: John Darrington 2004 +Copyright (C) 1997-9, 2000, 2007 Free Software Foundation, Inc. This program is free software; you can redistribute it and/or modify it under the terms of the GNU General Public License as @@ -26,7 +25,8 @@ Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA #include #include -#include +#include +#include #include #include #include @@ -39,9 +39,9 @@ Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA #include #include #include -#include #include #include +#include #include #include #include @@ -57,29 +57,24 @@ Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA /* (specification) "ONEWAY" (oneway_): *^variables=custom; - +missing=miss:!analysis/listwise, - incl:include/!exclude; - contrast= double list; - statistics[st_]=descriptives,homogeneity. + missing=miss:!analysis/listwise, + incl:include/!exclude; + +contrast= double list; + +statistics[st_]=descriptives,homogeneity. */ /* (declarations) */ /* (functions) */ - - -static int bad_weight_warn = 1; - - static struct cmd_oneway cmd; /* The independent variable */ -static struct variable *indep_var; +static const struct variable *indep_var; /* Number of dependent variables */ static size_t n_vars; /* The dependent variables */ -static struct variable **vars; +static const struct variable **vars; /* A hash table containing all the distinct values of the independent @@ -88,15 +83,11 @@ static struct hsh_table *global_group_hash ; /* The number of distinct values of the independent variable, when all missing values are disregarded */ -static int ostensible_number_of_groups=-1; +static int ostensible_number_of_groups = -1; -/* Function to use for testing for missing values */ -static is_missing_func *value_is_missing; - - -static bool run_oneway(const struct ccase *first, - const struct casefile *cf, void *_mode); +static void run_oneway (struct cmd_oneway *, struct casereader *, + const struct dataset *); /* Routines to show the output tables */ @@ -116,20 +107,16 @@ void output_oneway(void); int -cmd_oneway(void) +cmd_oneway (struct lexer *lexer, struct dataset *ds) { + struct casegrouper *grouper; + struct casereader *group; int i; bool ok; - if ( !parse_oneway(&cmd, NULL) ) + if ( !parse_oneway (lexer, ds, &cmd, NULL) ) return CMD_FAILURE; - /* If /MISSING=INCLUDE is set, then user missing values are ignored */ - if (cmd.incl == ONEWAY_INCLUDE ) - value_is_missing = mv_is_value_system_missing; - else - value_is_missing = mv_is_value_missing; - /* What statistics were requested */ if ( cmd.sbc_statistics ) { @@ -149,7 +136,12 @@ cmd_oneway(void) } } - ok = multipass_procedure_with_splits (run_oneway, &cmd); + /* Data pass. FIXME: error handling. */ + grouper = casegrouper_create_splits (proc_open (ds), dataset_dict (ds)); + while (casegrouper_get_next_group (grouper, &group)) + run_oneway (&cmd, group, ds); + ok = casegrouper_destroy (grouper); + ok = proc_commit (ds) && ok; free (vars); free_oneway (&cmd); @@ -186,7 +178,8 @@ output_oneway(void) sum += subc_list_double_at(&cmd.dl_contrast[i],j); if ( sum != 0.0 ) - msg(SW,_("Coefficients for contrast %d do not total zero"),i + 1); + msg(SW,_("Coefficients for contrast %d do not total zero"), + (int) i + 1); } if ( stat_tables & STAT_DESC ) @@ -223,17 +216,19 @@ output_oneway(void) /* Parser for the variables sub command */ static int -oneway_custom_variables(struct cmd_oneway *cmd UNUSED, void *aux UNUSED) +oneway_custom_variables (struct lexer *lexer, + struct dataset *ds, struct cmd_oneway *cmd UNUSED, + void *aux UNUSED) { + struct dictionary *dict = dataset_dict (ds); - lex_match('='); + lex_match (lexer, '='); - if ((token != T_ID || dict_lookup_var (default_dict, tokid) == NULL) - && token != T_ALL) + if ((lex_token (lexer) != T_ID || dict_lookup_var (dict, lex_tokid (lexer)) == NULL) + && lex_token (lexer) != T_ALL) return 2; - - if (!parse_variables (default_dict, &vars, &n_vars, + if (!parse_variables_const (lexer, dict, &vars, &n_vars, PV_DUPLICATE | PV_NUMERIC | PV_NO_SCRATCH) ) { @@ -243,19 +238,17 @@ oneway_custom_variables(struct cmd_oneway *cmd UNUSED, void *aux UNUSED) assert(n_vars); - if ( ! lex_match(T_BY)) + if ( ! lex_match (lexer, T_BY)) return 2; - - indep_var = parse_variable(); + indep_var = parse_variable (lexer, dict); if ( !indep_var ) { - msg(SE,_("`%s' is not a variable name"),tokid); + msg(SE,_("`%s' is not a variable name"),lex_tokid (lexer)); return 0; } - return 1; } @@ -362,10 +355,9 @@ show_anova_table(void) tab_title (t, _("ANOVA")); tab_submit (t); - - } + /* Show the descriptives table */ static void show_descriptives(void) @@ -446,7 +438,8 @@ show_descriptives(void) gs = gs_array[count]; tab_text (t, 1, row + count, - TAB_LEFT | TAT_TITLE ,value_to_string(&gs->id,indep_var)); + TAB_LEFT | TAT_TITLE, var_get_value_name(indep_var, + &gs->id)); /* Now fill in the numbers ... */ @@ -572,14 +565,12 @@ show_homogeneity(void) } tab_submit (t); - - } /* Show the contrast coefficients table */ static void -show_contrast_coeffs(short *bad_contrast) +show_contrast_coeffs (short *bad_contrast) { int n_cols = 2 + ostensible_number_of_groups; int n_rows = 2 + cmd.sbc_contrast; @@ -634,7 +625,7 @@ show_contrast_coeffs(short *bad_contrast) group_value = group_values[count]; tab_text (t, count + 2, 1, TAB_CENTER | TAT_TITLE, - value_to_string(group_value, indep_var)); + var_get_value_name (indep_var, group_value)); for (i = 0 ; i < cmd.sbc_contrast ; ++i ) { @@ -880,7 +871,7 @@ precalc ( struct cmd_oneway *cmd UNUSED ) (hsh_compare_func *) compare_group, (hsh_hash_func *) hash_group, (hsh_free_func *) free_group, - (void *) indep_var->width ); + (void *) var_get_width (indep_var) ); totals->sum=0; @@ -892,63 +883,64 @@ precalc ( struct cmd_oneway *cmd UNUSED ) } } +static void +free_value (void *value_, const void *aux UNUSED) +{ + union value *value = value_; + free (value); +} -static bool -run_oneway(const struct ccase *first, const struct casefile *cf, void *cmd_) +static void +run_oneway (struct cmd_oneway *cmd, + struct casereader *input, + const struct dataset *ds) { - struct casereader *r; + struct taint *taint; + struct dictionary *dict = dataset_dict (ds); + enum mv_class exclude; + struct casereader *reader; struct ccase c; - struct cmd_oneway *cmd = (struct cmd_oneway *) cmd_; + if (!casereader_peek (input, 0, &c)) + return; + output_split_file_values (ds, &c); + case_destroy (&c); - output_split_file_values (first); + taint = taint_clone (casereader_get_taint (input)); global_group_hash = hsh_create(4, (hsh_compare_func *) compare_values, (hsh_hash_func *) hash_value, - 0, - (void *) indep_var->width ); + free_value, + (void *) var_get_width (indep_var) ); + precalc(cmd); - for(r = casefile_get_reader (cf); - casereader_read (r, &c) ; - case_destroy (&c)) + exclude = cmd->incl != ONEWAY_INCLUDE ? MV_ANY : MV_SYSTEM; + input = casereader_create_filter_missing (input, &indep_var, 1, + exclude, NULL); + if (cmd->miss == ONEWAY_LISTWISE) + input = casereader_create_filter_missing (input, vars, n_vars, + exclude, NULL); + input = casereader_create_filter_weight (input, dict, NULL, NULL); + + reader = casereader_clone (input); + for (; casereader_read (reader, &c); case_destroy (&c)) { size_t i; - const double weight = - dict_get_case_weight(default_dict,&c,&bad_weight_warn); - - const union value *indep_val = case_data (&c, indep_var->fv); - - /* Deal with missing values */ - if ( value_is_missing(&indep_var->miss, indep_val) ) - continue; - - /* Skip the entire case if /MISSING=LISTWISE is set */ - if ( cmd->miss == ONEWAY_LISTWISE ) - { - for(i = 0; i < n_vars ; ++i) - { - const struct variable *v = vars[i]; - const union value *val = case_data (&c, v->fv); - - if (value_is_missing(&v->miss, val) ) - break; - } - if ( i != n_vars ) - continue; - - } + const double weight = dict_get_case_weight (dict, &c, NULL); - - hsh_insert ( global_group_hash, (void *) indep_val ); + const union value *indep_val = case_data (&c, indep_var); + void **p = hsh_probe (global_group_hash, indep_val); + if (*p == NULL) + *p = value_dup (indep_val, var_get_width (indep_var)); for ( i = 0 ; i < n_vars ; ++i ) { const struct variable *v = vars[i]; - const union value *val = case_data (&c, v->fv); + const union value *val = case_data (&c, v); struct group_proc *gp = group_proc_get (vars[i]); struct hsh_table *group_hash = gp->group_hash; @@ -970,8 +962,8 @@ run_oneway(const struct ccase *first, const struct casefile *cf, void *cmd_) hsh_insert ( group_hash, (void *) gs ); } - - if (! value_is_missing(&v->miss, val) ) + + if (!var_is_value_missing (v, val, exclude)) { struct group_statistics *totals = &gp->ugs; @@ -1000,22 +992,21 @@ run_oneway(const struct ccase *first, const struct casefile *cf, void *cmd_) } } - casereader_destroy (r); + casereader_destroy (reader); postcalc(cmd); if ( stat_tables & STAT_HOMO ) - levene(cf, indep_var, n_vars, vars, - (cmd->miss == ONEWAY_LISTWISE) ? LEV_LISTWISE : LEV_ANALYSIS , - value_is_missing); - - ostensible_number_of_groups = hsh_count (global_group_hash); + levene (dict, casereader_clone (input), indep_var, n_vars, vars, exclude); + casereader_destroy (input); - output_oneway(); + ostensible_number_of_groups = hsh_count (global_group_hash); - return true; + if (!taint_has_tainted_successor (taint)) + output_oneway(); + taint_destroy (taint); }