X-Git-Url: https://pintos-os.org/cgi-bin/gitweb.cgi?a=blobdiff_plain;f=src%2Flanguage%2Fstats%2Fkruskal-wallis.c;h=65ce7865a3f83ba9baea594756f2ca80aa32e4b3;hb=refs%2Fheads%2Fcenter-titles;hp=51ee291f95651be59c024cf4c57c49c8528e3769;hpb=471d0e518a94de5305b4697f25641464761ef910;p=pspp

diff --git a/src/language/stats/kruskal-wallis.c b/src/language/stats/kruskal-wallis.c
index 51ee291f95..65ce7865a3 100644
--- a/src/language/stats/kruskal-wallis.c
+++ b/src/language/stats/kruskal-wallis.c
@@ -1,5 +1,5 @@
 /* Pspp - a program for statistical analysis.
-   Copyright (C) 2010 Free Software Foundation, Inc.
+   Copyright (C) 2010, 2011 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 published by
@@ -22,25 +22,26 @@
 #include <gsl/gsl_cdf.h>
 #include <math.h>
 
-#include <data/casereader.h>
-#include <data/casewriter.h>
-#include <data/dictionary.h>
-#include <data/format.h>
-#include <data/procedure.h>
-#include <data/subcase.h>
-#include <data/variable.h>
-
-#include <libpspp/assertion.h>
-#include <libpspp/message.h>
-#include <libpspp/misc.h>
-#include <libpspp/hmap.h>
-#include <math/sort.h>
-
-
-#include "minmax.h"
-#include "xalloc.h"
-
-
+#include "data/casereader.h"
+#include "data/casewriter.h"
+#include "data/dataset.h"
+#include "data/dictionary.h"
+#include "data/format.h"
+#include "data/subcase.h"
+#include "data/variable.h"
+#include "libpspp/assertion.h"
+#include "libpspp/hmap.h"
+#include "libpspp/bt.h"
+#include "libpspp/message.h"
+#include "libpspp/misc.h"
+#include "math/sort.h"
+#include "output/tab.h"
+
+#include "gl/minmax.h"
+#include "gl/xalloc.h"
+
+
+/* Returns true iff the independent variable lies in the range [nst->val1, nst->val2] */
 static bool
 include_func (const struct ccase *c, void *aux)
 {
@@ -59,12 +60,28 @@ include_func (const struct ccase *c, void *aux)
 struct rank_entry
 {
   struct hmap_node node;
+  struct bt_node btn;
   union value group;
 
   double sum_of_ranks;
   double n;
 };
 
+
+static int
+compare_rank_entries_3way (const struct bt_node *a,
+                           const struct bt_node *b,
+                           const void *aux)
+{
+  const struct variable *var = aux;
+  const struct rank_entry *rea = bt_data (a, struct rank_entry, btn);
+  const struct rank_entry *reb = bt_data (b, struct rank_entry, btn);
+
+  return value_compare_3way (&rea->group, &reb->group, var_get_width (var));
+}
+
+
+/* Return the entry with the key GROUP or null if there is no such entry */
 static struct rank_entry *
 find_rank_entry (const struct hmap *map, const union value *group, size_t width)
 {
@@ -80,6 +97,7 @@ find_rank_entry (const struct hmap *map, const union value *group, size_t width)
   return re;
 }
 
+/* Calculates the adjustment necessary for tie compensation */
 static void
 distinct_callback (double v UNUSED, casenumber t, double w UNUSED, void *aux)
 {
@@ -181,6 +199,7 @@ kruskal_wallis_execute (const struct dataset *ds,
 
       casereader_destroy (rr);
 
+      /* Calculate the value of h */
       {
 	struct rank_entry *mre;
 	double n = 0.0;
@@ -204,6 +223,7 @@ kruskal_wallis_execute (const struct dataset *ds,
   show_ranks_box (nst, kw, total_n_groups);
   show_sig_box (nst, kw);
 
+  /* Cleanup allocated memory */
   for (i = 0 ; i < nst->n_vars; ++i)
     {
       struct rank_entry *mre, *next;
@@ -219,7 +239,6 @@ kruskal_wallis_execute (const struct dataset *ds,
 }
 
 
-#include <output/tab.h>
 #include "gettext.h"
 #define _(msgid) gettext (msgid)
 
@@ -261,7 +280,9 @@ show_ranks_box (const struct n_sample_test *nst, const struct kw *kw, int n_grou
   for (i = 0 ; i < nst->n_vars ; ++i)
     {
       int tot = 0;
-      const struct rank_entry *re;
+      struct rank_entry *re_x;
+      struct bt_node *bt_n = NULL;
+      struct bt bt;
 
       if (i > 0)
 	tab_hline (table, TAL_1, 0, tab_nc (table) -1, row);
@@ -269,23 +290,38 @@ show_ranks_box (const struct n_sample_test *nst, const struct kw *kw, int n_grou
       tab_text (table,  0, row,
 		TAT_TITLE, var_to_string (nst->vars[i]));
 
-      HMAP_FOR_EACH (re, const struct rank_entry, node, &kw[i].map)
+      /* Sort the rank entries, by iteratin the hash and putting the entries
+         into a binary tree. */
+      bt_init (&bt, compare_rank_entries_3way, nst->vars[i]);
+      HMAP_FOR_EACH (re_x, struct rank_entry, node, &kw[i].map)
 	{
+          bt_insert (&bt, &re_x->btn);
+        }
+
+      /* Report the rank entries in sorted order. */
+      for (bt_n = bt_first (&bt);
+           bt_n != NULL;
+           bt_n = bt_next (&bt, bt_n) )
+        {
+          const struct rank_entry *re =
+            bt_data (bt_n, const struct rank_entry, btn);
+
 	  struct string str;
 	  ds_init_empty (&str);
-
+          
 	  var_append_value_name (nst->indep_var, &re->group, &str);
-
+          
 	  tab_text   (table, 1, row, TAB_LEFT, ds_cstr (&str));
-	  tab_double (table, 2, row, TAB_LEFT, re->n, &F_8_0);
-	  tab_double (table, 3, row, TAB_LEFT, re->sum_of_ranks / re->n, 0);
-
+	  tab_double (table, 2, row, TAB_LEFT, re->n, NULL, RC_INTEGER);
+	  tab_double (table, 3, row, TAB_LEFT, re->sum_of_ranks / re->n, NULL, RC_OTHER);
+          
 	  tot += re->n;
 	  row++;
 	  ds_destroy (&str);
 	}
+
       tab_double (table, 2, row, TAB_LEFT,
-		  tot, &F_8_0);
+		  tot, NULL, RC_INTEGER);
       tab_text (table, 1, row++, TAB_LEFT, _("Total"));
     }
 
@@ -331,14 +367,14 @@ show_sig_box (const struct n_sample_test *nst, const struct kw *kw)
 		);
 
       tab_double (table, column_headers + 1 + i, 1, 0,
-		  kw[i].h, 0);
+		  kw[i].h, NULL, RC_OTHER);
 
       tab_double (table, column_headers + 1 + i, 2, 0,
-		  df, &F_8_0);
+		  df, NULL, RC_INTEGER);
 
       tab_double (table, column_headers + 1 + i, 3, 0,
 		  gsl_cdf_chisq_Q (kw[i].h, df),
-		  0);
+		  NULL, RC_PVALUE);
     }
 
   tab_submit (table);