Merge commit 'origin/stable'

[pspp-builds.git] / src / data / category.c

diff --git a/src/data/category.c b/src/data/category.c
index b0424d9f65c054df4631897702d33ab4eae1c57a..968dd4c52da8369e629fb2797faa056db4b67227 100644 (file)
--- a/src/data/category.c
+++ b/src/data/category.c
@@ -1,5 +1,5 @@
 /* PSPP - a program for statistical analysis.
-   Copyright (C) 2005 Free Software Foundation, Inc.
+   Copyright (C) 2005, 2009 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
@@ -31,17 +31,15 @@
 #include <config.h>
 
 #include <assert.h>
+#include <data/category.h>
+#include <data/value.h>
+#include <data/variable.h>
+#include <gl/xalloc.h>
+#include <libpspp/message.h>
 #include <stdlib.h>
 #include <string.h>
 
-#include <libpspp/message.h>
-#include "category.h"
-#include "value.h"
-#include "variable.h"
-
-#include "xalloc.h"
-
-#define CAT_VALUE_NOT_FOUND -2
+#define CAT_VALUE_NOT_FOUND -1
 
 #define N_INITIAL_CATEGORIES 1
 
@@ -58,6 +56,11 @@ struct cat_vals
                                           track of the number of
                                           values stored.
                                         */
+  size_t *value_counts; /* Element i stores the number of cases for which
+                          the categorical variable has that corresponding 
+                          value. This is necessary for computing covariance
+                          matrices.
+                        */
 };
 
 void
@@ -70,6 +73,7 @@ cat_stored_values_create (const struct variable *v)
       obs_vals->n_categories = 0;
       obs_vals->n_allocated_categories = N_INITIAL_CATEGORIES;
       obs_vals->vals = xnmalloc (N_INITIAL_CATEGORIES, sizeof *obs_vals->vals);
+      obs_vals->value_counts = xnmalloc (N_INITIAL_CATEGORIES, sizeof *obs_vals->value_counts);
       var_set_obs_vals (v, obs_vals);
     }
 }
@@ -80,7 +84,10 @@ cat_stored_values_destroy (struct cat_vals *obs_vals)
   if (obs_vals != NULL)
     {
       if (obs_vals->n_allocated_categories > 0)
-        free (obs_vals->vals);
+       {
+         free (obs_vals->vals);
+         free (obs_vals->value_counts);
+       }
       free (obs_vals);
     }
 }
@@ -99,7 +106,7 @@ cat_value_find (const struct variable *v, const union value *val)
     {
       candidate = obs_vals->vals + i;
       assert (candidate != NULL);
-      if (!compare_values (candidate, val, var_get_width (v)))
+      if (value_equal (candidate, val, var_get_width (v)))
        {
          return i;
        }
@@ -108,15 +115,17 @@ cat_value_find (const struct variable *v, const union value *val)
 }
 
 /*
-   Add the new value unless it is already present.
+   Add the new value unless it is already present. Increment the count.
  */
 void
 cat_value_update (const struct variable *v, const union value *val)
 {
   if (var_is_alpha (v))
     {
+      size_t i;
       struct cat_vals *cv = var_get_obs_vals (v);
-      if (cat_value_find (v, val) == CAT_VALUE_NOT_FOUND)
+      i = cat_value_find (v, val);
+      if (i == CAT_VALUE_NOT_FOUND)
        {
          if (cv->n_categories >= cv->n_allocated_categories)
            {
@@ -124,11 +133,35 @@ cat_value_update (const struct variable *v, const union value *val)
              cv->vals = xnrealloc (cv->vals,
                                    cv->n_allocated_categories,
                                    sizeof *cv->vals);
+             cv->value_counts = xnrealloc (cv->value_counts, cv->n_allocated_categories,
+                                           sizeof *cv->value_counts);
            }
          cv->vals[cv->n_categories] = *val;
+         cv->value_counts[cv->n_categories] = 1;
          cv->n_categories++;
        }
+      else
+       {
+         cv->value_counts[i]++;
+       }
+    }
+}
+/*
+  Return the count for the sth category.
+ */
+size_t
+cat_get_category_count (const size_t s, const struct variable *v)
+{
+  struct cat_vals *tmp;
+  size_t n_categories;
+
+  tmp = var_get_obs_vals (v);
+  n_categories = cat_get_n_categories (v);
+  if (s < n_categories)
+    {
+      return tmp->value_counts[s];
     }
+  return CAT_VALUE_NOT_FOUND;
 }
 
 const union value *
@@ -147,3 +180,20 @@ cat_get_n_categories (const struct variable *v)
   return var_get_obs_vals (v)->n_categories;
 }
 
+/*
+  If VAR is categorical with d categories, its first category should
+  correspond to the origin in d-dimensional Euclidean space.
+ */
+bool
+cat_is_origin (const struct variable *var, const union value *val)
+{
+  if (var_is_numeric (var))
+    {
+      return false;
+    }
+  if (cat_value_find (var, val) == 0)
+    {
+      return true;
+    }
+  return false;
+}