X-Git-Url: https://pintos-os.org/cgi-bin/gitweb.cgi?a=blobdiff_plain;f=src%2Flanguage%2Fstats%2Ffrequencies.q;h=ef4b7f9589d3a770fc9650a1a93b5581e9a2e33e;hb=9ade26c8349b4434008c46cf09bc7473ec743972;hp=ba7dd57feb35a13c56b3d3920e9ada048e4dbb0f;hpb=74242052a357f0285b5ae7d02a3bef4cd2017c7d;p=pspp-builds.git

diff --git a/src/language/stats/frequencies.q b/src/language/stats/frequencies.q
index ba7dd57f..ef4b7f95 100644
--- a/src/language/stats/frequencies.q
+++ b/src/language/stats/frequencies.q
@@ -1,5 +1,5 @@
 /* PSPP - a program for statistical analysis.
-   Copyright (C) 1997-9, 2000, 2007, 2009, 2010 Free Software Foundation, Inc.
+   Copyright (C) 1997-9, 2000, 2007, 2009, 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
@@ -125,10 +125,6 @@ struct percentile
 {
   double p;        /* the %ile to be calculated */
   double value;    /* the %ile's value */
-  double x1;       /* The datum value <= the percentile */
-  double x2;       /* The datum value >= the percentile */
-  int flag;
-  int flag2;       /* Set to 1 if this percentile value has been found */
   bool show;       /* True to show this percentile in the statistics box. */
 };
 
@@ -222,8 +218,9 @@ struct frq_proc
 static void determine_charts (struct frq_proc *,
                               const struct cmd_frequencies *);
 
-static void calc_stats (const struct frq_proc *, const struct var_freqs *,
-                        double d[FRQ_N_STATS]);
+static void calc_stats (const struct var_freqs *, double d[FRQ_N_STATS]);
+static void calc_percentiles (const struct frq_proc *,
+                              const struct var_freqs *);
 
 static void precalc (struct frq_proc *, struct casereader *, struct dataset *);
 static void calc (struct frq_proc *, const struct ccase *,
@@ -237,11 +234,6 @@ static void dump_statistics (const struct frq_proc *, const struct var_freqs *,
                              const struct variable *weight_var);
 static void cleanup_freq_tab (struct var_freqs *);
 
-static algo_compare_func compare_value_numeric_a, compare_value_alpha_a;
-static algo_compare_func compare_value_numeric_d, compare_value_alpha_d;
-static algo_compare_func compare_freq_numeric_a, compare_freq_alpha_a;
-static algo_compare_func compare_freq_numeric_d, compare_freq_alpha_d;
-
 static void add_percentile (struct frq_proc *, double x, bool show,
                             size_t *allocated_percentiles);
 
@@ -502,7 +494,7 @@ postcalc (struct frq_proc *frq, const struct dataset *ds)
 	  double d[FRQ_N_STATS];
 	  struct histogram *histogram;
 
-	  calc_stats (frq, vf, d);
+	  calc_stats (vf, d);
 
 	  histogram = freq_tab_to_hist (frq, &vf->tab, vf->var);
 
@@ -524,28 +516,6 @@ postcalc (struct frq_proc *frq, const struct dataset *ds)
     }
 }
 
-/* Returns the comparison function that should be used for
-   sorting a frequency table by FRQ_SORT using VAL_TYPE
-   values. */
-static algo_compare_func *
-get_freq_comparator (int frq_sort, enum val_type val_type)
-{
-  bool is_numeric = val_type == VAL_NUMERIC;
-  switch (frq_sort)
-    {
-    case FRQ_AVALUE:
-      return is_numeric ? compare_value_numeric_a : compare_value_alpha_a;
-    case FRQ_DVALUE:
-      return is_numeric ? compare_value_numeric_d : compare_value_alpha_d;
-    case FRQ_AFREQ:
-      return is_numeric ? compare_freq_numeric_a : compare_freq_alpha_a;
-    case FRQ_DFREQ:
-      return is_numeric ? compare_freq_numeric_d : compare_freq_alpha_d;
-    default:
-      NOT_REACHED ();
-    }
-}
-
 /* Returns true iff the value in struct freq F is non-missing
    for variable V. */
 static bool
@@ -557,12 +527,39 @@ not_missing (const void *f_, const void *v_)
   return !var_is_value_missing (v, &f->value, MV_ANY);
 }
 
+struct freq_compare_aux
+  {
+    bool by_freq;
+    bool ascending_freq;
+
+    int width;
+    bool ascending_value;
+  };
+
+static int
+compare_freq (const void *a_, const void *b_, const void *aux_)
+{
+  const struct freq_compare_aux *aux = aux_;
+  const struct freq *a = a_;
+  const struct freq *b = b_;
+
+  if (aux->by_freq && a->count != b->count)
+    {
+      int cmp = a->count > b->count ? 1 : -1;
+      return aux->ascending_freq ? cmp : -cmp;
+    }
+  else
+    {
+      int cmp = value_compare_3way (&a->value, &b->value, aux->width);
+      return aux->ascending_value ? cmp : -cmp;
+    }
+}
 /* Summarizes the frequency table data for variable V. */
 static void
 postprocess_freq_tab (const struct frq_proc *frq, struct var_freqs *vf)
 {
   struct freq_tab *ft = &vf->tab;
-  algo_compare_func *compare;
+  struct freq_compare_aux aux;
   size_t count;
   struct freq *freqs, *f;
   size_t i;
@@ -578,9 +575,12 @@ postprocess_freq_tab (const struct frq_proc *frq, struct var_freqs *vf)
   ft->n_missing = count - ft->n_valid;
 
   /* Sort data. */
-  compare = get_freq_comparator (frq->sort, var_get_type (vf->var));
-  sort (ft->valid, ft->n_valid, sizeof *ft->valid, compare, vf);
-  sort (ft->missing, ft->n_missing, sizeof *ft->missing, compare, vf);
+  aux.by_freq = frq->sort == FRQ_AFREQ || frq->sort == FRQ_DFREQ;
+  aux.ascending_freq = frq->sort != FRQ_DFREQ;
+  aux.width = vf->width;
+  aux.ascending_value = frq->sort != FRQ_DVALUE;
+  sort (ft->valid, ft->n_valid, sizeof *ft->valid, compare_freq, &aux);
+  sort (ft->missing, ft->n_missing, sizeof *ft->missing, compare_freq, &aux);
 
   /* Summary statistics. */
   ft->valid_cases = 0.0;
@@ -618,10 +618,10 @@ frq_custom_variables (struct lexer *lexer, struct dataset *ds,
   size_t n_vars;
   size_t i;
 
-  lex_match (lexer, '=');
+  lex_match (lexer, T_EQUALS);
   if (lex_token (lexer) != T_ALL
       && (lex_token (lexer) != T_ID
-          || dict_lookup_var (dataset_dict (ds), lex_tokid (lexer)) == NULL))
+          || dict_lookup_var (dataset_dict (ds), lex_tokcstr (lexer)) == NULL))
     return 2;
 
   /* Get list of current variables, to avoid duplicates. */
@@ -662,8 +662,9 @@ frq_custom_grouped (struct lexer *lexer, struct dataset *ds, struct cmd_frequenc
 {
   struct frq_proc *frq = frq_;
 
-  lex_match (lexer, '=');
-  if ((lex_token (lexer) == T_ID && dict_lookup_var (dataset_dict (ds), lex_tokid (lexer)) != NULL)
+  lex_match (lexer, T_EQUALS);
+  if ((lex_token (lexer) == T_ID
+       && dict_lookup_var (dataset_dict (ds), lex_tokcstr (lexer)) != NULL)
       || lex_token (lexer) == T_ID)
     for (;;)
       {
@@ -680,7 +681,7 @@ frq_custom_grouped (struct lexer *lexer, struct dataset *ds, struct cmd_frequenc
 	if (!parse_variables_const (lexer, dataset_dict (ds), &v, &n,
                               PV_NO_DUPLICATE | PV_NUMERIC))
 	  return 0;
-	if (lex_match (lexer, '('))
+	if (lex_match (lexer, T_LPAREN))
 	  {
 	    nl = ml = 0;
 	    dl = NULL;
@@ -693,11 +694,11 @@ frq_custom_grouped (struct lexer *lexer, struct dataset *ds, struct cmd_frequenc
 		  }
 		dl[nl++] = lex_tokval (lexer);
 		lex_get (lexer);
-		lex_match (lexer, ',');
+		lex_match (lexer, T_COMMA);
 	      }
 	    /* Note that nl might still be 0 and dl might still be
 	       NULL.  That's okay. */
-	    if (!lex_match (lexer, ')'))
+	    if (!lex_match (lexer, T_RPAREN))
 	      {
 		free (v);
 		msg (SE, _("`)' expected after GROUPED interval list."));
@@ -737,14 +738,23 @@ frq_custom_grouped (struct lexer *lexer, struct dataset *ds, struct cmd_frequenc
           }
 
 	free (v);
-	if (!lex_match (lexer, '/'))
-	  break;
-	if ((lex_token (lexer) != T_ID || dict_lookup_var (dataset_dict (ds), lex_tokid (lexer)) != NULL)
-            && lex_token (lexer) != T_ALL)
-	  {
-	    lex_put_back (lexer, '/');
-	    break;
-	  }
+        if (lex_token (lexer) != T_SLASH)
+          break;
+
+        if ((lex_next_token (lexer, 1) == T_ID
+             && dict_lookup_var (dataset_dict (ds),
+                                 lex_next_tokcstr (lexer, 1)))
+            || lex_next_token (lexer, 1) == T_ALL)
+          {
+            /* The token after the slash is a variable name.  Keep parsing. */
+            lex_get (lexer);
+          }
+        else
+          {
+            /* The token after the slash must be the start of a new
+               subcommand.  Let the caller see the slash. */
+            break;
+          }
       }
 
   return 1;
@@ -792,122 +802,6 @@ add_percentile (struct frq_proc *frq, double x, bool show,
 
 /* Comparison functions. */
 
-/* Ascending numeric compare of values. */
-static int
-compare_value_numeric_a (const void *a_, const void *b_,
-                         const void *vf_ UNUSED)
-{
-  const struct freq *a = a_;
-  const struct freq *b = b_;
-
-  if (a->value.f > b->value.f)
-    return 1;
-  else if (a->value.f < b->value.f)
-    return -1;
-  else
-    return 0;
-}
-
-/* Ascending string compare of values. */
-static int
-compare_value_alpha_a (const void *a_, const void *b_, const void *vf_)
-{
-  const struct freq *a = a_;
-  const struct freq *b = b_;
-  const struct var_freqs *vf = vf_;
-
-  return value_compare_3way (&a->value, &b->value, vf->width);
-}
-
-/* Descending numeric compare of values. */
-static int
-compare_value_numeric_d (const void *a, const void *b, const void *vf_ UNUSED)
-{
-  return -compare_value_numeric_a (a, b, vf_);
-}
-
-/* Descending string compare of values. */
-static int
-compare_value_alpha_d (const void *a, const void *b, const void *vf_)
-{
-  return -compare_value_alpha_a (a, b, vf_);
-}
-
-/* Ascending numeric compare of frequency;
-   secondary key on ascending numeric value. */
-static int
-compare_freq_numeric_a (const void *a_, const void *b_, const void *vf_ UNUSED)
-{
-  const struct freq *a = a_;
-  const struct freq *b = b_;
-
-  if (a->count > b->count)
-    return 1;
-  else if (a->count < b->count)
-    return -1;
-
-  if (a->value.f > b->value.f)
-    return 1;
-  else if (a->value.f < b->value.f)
-    return -1;
-  else
-    return 0;
-}
-
-/* Ascending numeric compare of frequency;
-   secondary key on ascending string value. */
-static int
-compare_freq_alpha_a (const void *a_, const void *b_, const void *vf_)
-{
-  const struct freq *a = a_;
-  const struct freq *b = b_;
-  const struct var_freqs *vf = vf_;
-
-  if (a->count > b->count)
-    return 1;
-  else if (a->count < b->count)
-    return -1;
-  else
-    return value_compare_3way (&a->value, &b->value, vf->width);
-}
-
-/* Descending numeric compare of frequency;
-   secondary key on ascending numeric value. */
-static int
-compare_freq_numeric_d (const void *a_, const void *b_, const void *vf_ UNUSED)
-{
-  const struct freq *a = a_;
-  const struct freq *b = b_;
-
-  if (a->count > b->count)
-    return -1;
-  else if (a->count < b->count)
-    return 1;
-
-  if (a->value.f > b->value.f)
-    return 1;
-  else if (a->value.f < b->value.f)
-    return -1;
-  else
-    return 0;
-}
-
-/* Descending numeric compare of frequency;
-   secondary key on ascending string value. */
-static int
-compare_freq_alpha_d (const void *a_, const void *b_, const void *vf_)
-{
-  const struct freq *a = a_;
-  const struct freq *b = b_;
-  const struct var_freqs *vf = vf_;
-
-  if (a->count > b->count)
-    return -1;
-  else if (a->count < b->count)
-    return 1;
-  else
-    return value_compare_3way (&a->value, &b->value, vf->width);
-}
 
 /* Frequency table display. */
 
@@ -996,114 +890,73 @@ dump_freq_table (const struct var_freqs *vf, const struct variable *wv)
 
 /* Statistical display. */
 
-/* Calculates all the pertinent statistics for variable V, putting them in
-   array D[]. */
+static double
+calc_percentile (double p, double valid_cases, double x1, double x2)
+{
+  double s, dummy;
+
+  s = (settings_get_algorithm () != COMPATIBLE
+       ? modf ((valid_cases - 1) * p, &dummy)
+       : modf ((valid_cases + 1) * p - 1, &dummy));
+
+  return x1 + (x2 - x1) * s;
+}
+
+/* Calculates all of the percentiles for VF within FRQ. */
 static void
-calc_stats (const struct frq_proc *frq,
-            const struct var_freqs *vf, double d[FRQ_N_STATS])
+calc_percentiles (const struct frq_proc *frq, const struct var_freqs *vf)
 {
   const struct freq_tab *ft = &vf->tab;
   double W = ft->valid_cases;
-  struct moments *m;
-  struct freq *f=0;
-  double prev_value;
-  int most_often;
-  double X_mode;
-
+  const struct freq *f;
+  int percentile_idx;
   double rank;
-  int i = 0;
-  int idx;
-
-  /* Calculate percentiles. */
 
   assert (ft->n_valid > 0);
 
-  for (i = 0; i < frq->n_percentiles; i++)
-    {
-      struct percentile *pc = &frq->percentiles[i];
-
-      pc->flag = 0;
-      pc->flag2 = 0;
-    }
-
   rank = 0;
-  prev_value = SYSMIS;
-  for (idx = 0; idx < ft->n_valid; ++idx)
+  percentile_idx = 0;
+  for (f = ft->valid; f < ft->missing; f++)
     {
-      f = &ft->valid[idx];
-      rank += f->count ;
-      for (i = 0; i < frq->n_percentiles; i++)
+      rank += f->count;
+      for (; percentile_idx < frq->n_percentiles; percentile_idx++)
         {
-          struct percentile *pc = &frq->percentiles[i];
-	  double tp;
-
-	  if ( pc->flag2  ) continue ;
-
-	  if ( settings_get_algorithm () != COMPATIBLE )
-	    tp =
-	      (ft->valid_cases - 1) *  pc->p;
-	  else
-	    tp =
-	      (ft->valid_cases + 1) *  pc->p - 1;
-
-	  if ( pc->flag )
-	    {
-	      pc->x2 = f->value.f;
-	      pc->x1 = prev_value;
-	      pc->flag2 = 1;
-	      continue;
-	    }
-
-          if (rank >  tp )
-	  {
-	    if ( f->count > 1 && rank - (f->count - 1) > tp )
-	      {
-		pc->x2 = pc->x1 = f->value.f;
-		pc->flag2 = 1;
-	      }
-	    else
-	      {
-		pc->flag=1;
-	      }
+          struct percentile *pc = &frq->percentiles[percentile_idx];
+          double tp;
 
-	    continue;
-	  }
-        }
-      prev_value = f->value.f;
-    }
+          tp = (settings_get_algorithm () == ENHANCED
+                ? (W - 1) * pc->p
+                : (W + 1) * pc->p - 1);
 
-  for (i = 0; i < frq->n_percentiles; i++)
-    {
-      struct percentile *pc = &frq->percentiles[i];
-
-      /* Catches the case when p == 100% */
-      if ( ! pc->flag2 )
-	pc->x1 = pc->x2 = f->value.f;
+          if (rank <= tp)
+            break;
 
-      /*
-      printf("percentile %d (p==%.2f); X1 = %g; X2 = %g\n",
-	     i,pc->p,pc->x1,pc->x2);
-      */
+          if (tp + 1 < rank || f + 1 >= ft->missing)
+            pc->value = f->value.f;
+          else
+            pc->value = calc_percentile (pc->p, W, f->value.f, f[1].value.f);
+        }
     }
-
-  for (i = 0; i < frq->n_percentiles; i++)
+  for (; percentile_idx < frq->n_percentiles; percentile_idx++)
     {
-      struct percentile *pc = &frq->percentiles[i];
-      double s;
-
-      double dummy;
-      if ( settings_get_algorithm () != COMPATIBLE )
-	{
-	  s = modf((ft->valid_cases - 1) * pc->p , &dummy);
-	}
-      else
-	{
-	  s = modf((ft->valid_cases + 1) * pc->p -1, &dummy);
-	}
-
-      pc->value = pc->x1 + (pc->x2 - pc->x1) * s ;
+      struct percentile *pc = &frq->percentiles[percentile_idx];
+      pc->value = ft->valid[ft->n_valid - 1].value.f;
     }
+}
 
+/* Calculates all the pertinent statistics for VF, putting them in array
+   D[]. */
+static void
+calc_stats (const struct var_freqs *vf, double d[FRQ_N_STATS])
+{
+  const struct freq_tab *ft = &vf->tab;
+  double W = ft->valid_cases;
+  const struct freq *f;
+  struct moments *m;
+  int most_often;
+  double X_mode;
+
+  assert (ft->n_valid > 0);
 
   /* Calculate the mode. */
   most_often = -1;
@@ -1165,7 +1018,8 @@ dump_statistics (const struct frq_proc *frq, const struct var_freqs *vf,
 	   var_get_name (vf->var));
       return;
     }
-  calc_stats (frq, vf, stat_value);
+  calc_stats (vf, stat_value);
+  calc_percentiles (frq, vf);
 
   t = tab_create (3, frq->n_stats + frq->n_show_percentiles + 2);