X-Git-Url: https://pintos-os.org/cgi-bin/gitweb.cgi?a=blobdiff_plain;f=src%2Flanguage%2Fstats%2Ffrequencies.q;h=c00e645f82e9ccd27a05e4a540b2f96e813405e2;hb=2be9bee9da6a2ce27715e58128569594319abfa2;hp=2ede5b4728d54353645fcb490285560d87478019;hpb=87fb226f77978ca52736d7fb010c53f80f4ce4b9;p=pspp-builds.git

diff --git a/src/language/stats/frequencies.q b/src/language/stats/frequencies.q
index 2ede5b47..c00e645f 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
@@ -23,9 +23,9 @@
 #include "data/case.h"
 #include "data/casegrouper.h"
 #include "data/casereader.h"
+#include "data/dataset.h"
 #include "data/dictionary.h"
 #include "data/format.h"
-#include "data/procedure.h"
 #include "data/settings.h"
 #include "data/value-labels.h"
 #include "data/variable.h"
@@ -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. */
 };
 
@@ -143,10 +139,6 @@ struct percentile
 	 | BI(FRQ_SKEW) | BI(FRQ_SESKEW) | BI(FRQ_RANGE)	\
 	 | BI(FRQ_RANGE) | BI(FRQ_MODE) | BI(FRQ_MEDIAN))
 
-/* Statistics; number of statistics. */
-static unsigned long stats;
-static int n_stats;
-
 struct frq_chart
   {
     double x_min;               /* X axis minimum value. */
@@ -161,12 +153,6 @@ struct frq_chart
     bool include_missing;       /* Whether to include missing values. */
   };
 
-/* Histogram and pie chart settings. */
-static struct frq_chart hist, pie;
-
-/* Parsed command. */
-static struct cmd_frequencies cmd;
-
 /* Frequency tables. */
 
 /* Entire frequency table. */
@@ -212,37 +198,47 @@ struct frq_proc
     struct var_freqs *vars;
     size_t n_vars;
 
+    /* Percentiles to calculate and possibly display. */
     struct percentile *percentiles;
     int n_percentiles, n_show_percentiles;
+
+    /* Frequency table display. */
+    int max_categories;         /* Maximum categories to show. */
+    int sort;                   /* FRQ_AVALUE or FRQ_DVALUE
+                                   or FRQ_ACOUNT or FRQ_DCOUNT. */
+
+    /* Statistics; number of statistics. */
+    unsigned long stats;
+    int n_stats;
+
+    /* Histogram and pie chart settings. */
+    struct frq_chart *hist, *pie;
   };
 
-static void determine_charts (void);
+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 *,
                   const struct dataset *);
 static void postcalc (struct frq_proc *, const struct dataset *);
 
-static void postprocess_freq_tab (struct var_freqs *);
+static void postprocess_freq_tab (const struct frq_proc *, struct var_freqs *);
 static void dump_freq_table (const struct var_freqs *,
                              const struct variable *weight_var);
 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);
 
-static void do_piechart(const struct variable *var,
-			const struct freq_tab *frq_tab);
+static void do_piechart(const struct frq_chart *, const struct variable *,
+			const struct freq_tab *);
 
 struct histogram *freq_tab_to_hist(const struct frq_proc *,
                                    const struct freq_tab *,
@@ -253,6 +249,7 @@ struct histogram *freq_tab_to_hist(const struct frq_proc *,
 int
 cmd_frequencies (struct lexer *lexer, struct dataset *ds)
 {
+  struct cmd_frequencies cmd;
   struct frq_proc frq;
   struct casegrouper *grouper;
   struct casereader *input, *group;
@@ -269,6 +266,9 @@ cmd_frequencies (struct lexer *lexer, struct dataset *ds)
   frq.n_percentiles = 0;
   frq.n_show_percentiles = 0;
 
+  frq.hist = NULL;
+  frq.pie = NULL;
+
   allocated_percentiles = 0;
 
   if (!parse_frequencies (lexer, ds, &cmd, &frq))
@@ -277,30 +277,36 @@ cmd_frequencies (struct lexer *lexer, struct dataset *ds)
       return CMD_FAILURE;
     }
 
+  /* Figure out when to show frequency tables. */
+  frq.max_categories = (cmd.table == FRQ_NOTABLE ? -1
+                        : cmd.table == FRQ_TABLE ? INT_MAX
+                        : cmd.limit);
+  frq.sort = cmd.sort;
+
   /* Figure out statistics to calculate. */
-  stats = 0;
+  frq.stats = 0;
   if (cmd.a_statistics[FRQ_ST_DEFAULT] || !cmd.sbc_statistics)
-    stats |= FRQ_DEFAULT;
+    frq.stats |= FRQ_DEFAULT;
   if (cmd.a_statistics[FRQ_ST_ALL])
-    stats |= FRQ_ALL;
+    frq.stats |= FRQ_ALL;
   if (cmd.sort != FRQ_AVALUE && cmd.sort != FRQ_DVALUE)
-    stats &= ~BIT_INDEX (FRQ_MEDIAN);
+    frq.stats &= ~BIT_INDEX (FRQ_MEDIAN);
   for (i = 0; i < FRQ_N_STATS; i++)
     if (cmd.a_statistics[st_name[i].st_indx])
-      stats |= BIT_INDEX (i);
-  if (stats & FRQ_KURT)
-    stats |= BIT_INDEX (FRQ_SEKURT);
-  if (stats & FRQ_SKEW)
-    stats |= BIT_INDEX (FRQ_SESKEW);
+      frq.stats |= BIT_INDEX (i);
+  if (frq.stats & FRQ_KURT)
+    frq.stats |= BIT_INDEX (FRQ_SEKURT);
+  if (frq.stats & FRQ_SKEW)
+    frq.stats |= BIT_INDEX (FRQ_SESKEW);
 
   /* Calculate n_stats. */
-  n_stats = 0;
+  frq.n_stats = 0;
   for (i = 0; i < FRQ_N_STATS; i++)
-    if ((stats & BIT_INDEX (i)))
-      n_stats++;
+    if ((frq.stats & BIT_INDEX (i)))
+      frq.n_stats++;
 
   /* Charting. */
-  determine_charts ();
+  determine_charts (&frq, &cmd);
   if (cmd.sbc_histogram || cmd.sbc_piechart || cmd.sbc_ntiles)
     cmd.sort = FRQ_AVALUE;
 
@@ -326,13 +332,13 @@ cmd_frequencies (struct lexer *lexer, struct dataset *ds)
                             &allocated_percentiles);
 	}
     }
-  if (stats & BIT_INDEX (FRQ_MEDIAN))
+  if (frq.stats & BIT_INDEX (FRQ_MEDIAN))
     {
       /* Treat the median as the 50% percentile.
          We output it in the percentiles table as "50 (Median)." */
       add_percentile (&frq, 0.5, true, &allocated_percentiles);
-      stats &= ~BIT_INDEX (FRQ_MEDIAN);
-      n_stats--;
+      frq.stats &= ~BIT_INDEX (FRQ_MEDIAN);
+      frq.n_stats--;
     }
   if (cmd.sbc_histogram)
     {
@@ -362,53 +368,61 @@ cmd_frequencies (struct lexer *lexer, struct dataset *ds)
   pool_destroy (frq.pool);
   free (frq.vars);
   free (frq.percentiles);
+  free (frq.hist);
+  free (frq.pie);
 
   return ok ? CMD_SUCCESS : CMD_CASCADING_FAILURE;
 }
 
 /* Figure out which charts the user requested.  */
 static void
-determine_charts (void)
+determine_charts (struct frq_proc *frq, const struct cmd_frequencies *cmd)
 {
-  if (cmd.sbc_barchart)
+  if (cmd->sbc_barchart)
     msg (SW, _("Bar charts are not implemented."));
 
-  if (cmd.sbc_histogram)
+  if (cmd->sbc_histogram)
     {
-      hist.x_min = cmd.hi_min;
-      hist.x_max = cmd.hi_max;
-      hist.y_scale = cmd.hi_scale;
-      hist.y_max = cmd.hi_scale == FRQ_FREQ ? cmd.hi_freq : cmd.hi_pcnt;
-      hist.draw_normal = cmd.hi_norm != FRQ_NONORMAL;
-      hist.include_missing = false;
-
-      if (hist.x_min != SYSMIS && hist.x_max != SYSMIS
-          && hist.x_min >= hist.x_max)
+      struct frq_chart *hist;
+
+      hist = frq->hist = xmalloc (sizeof *frq->hist);
+      hist->x_min = cmd->hi_min;
+      hist->x_max = cmd->hi_max;
+      hist->y_scale = cmd->hi_scale;
+      hist->y_max = cmd->hi_scale == FRQ_FREQ ? cmd->hi_freq : cmd->hi_pcnt;
+      hist->draw_normal = cmd->hi_norm != FRQ_NONORMAL;
+      hist->include_missing = false;
+
+      if (hist->x_min != SYSMIS && hist->x_max != SYSMIS
+          && hist->x_min >= hist->x_max)
         {
           msg (SE, _("MAX for histogram must be greater than or equal to MIN, "
                      "but MIN was specified as %.15g and MAX as %.15g.  "
-                     "MIN and MAX will be ignored."), hist.x_min, hist.x_max);
-          hist.x_min = hist.x_max = SYSMIS;
+                     "MIN and MAX will be ignored."),
+               hist->x_min, hist->x_max);
+          hist->x_min = hist->x_max = SYSMIS;
         }
     }
 
-  if (cmd.sbc_piechart)
+  if (cmd->sbc_piechart)
     {
-      pie.x_min = cmd.pie_min;
-      pie.x_max = cmd.pie_max;
-      pie.y_scale = cmd.pie_scale;
-      pie.include_missing = cmd.pie_missing == FRQ_MISSING;
+      struct frq_chart *pie;
 
-      if (pie.x_min != SYSMIS && pie.x_max != SYSMIS
-          && pie.x_min >= pie.x_max)
+      pie = frq->pie = xmalloc (sizeof *frq->pie);
+      pie->x_min = cmd->pie_min;
+      pie->x_max = cmd->pie_max;
+      pie->y_scale = cmd->pie_scale;
+      pie->include_missing = cmd->pie_missing == FRQ_MISSING;
+
+      if (pie->x_min != SYSMIS && pie->x_max != SYSMIS
+          && pie->x_min >= pie->x_max)
         {
           msg (SE, _("MAX for pie chart must be greater than or equal to MIN, "
                      "but MIN was specified as %.15g and MAX as %.15g.  "
-                     "MIN and MAX will be ignored."), pie.x_min, pie.x_max);
-          pie.x_min = pie.x_max = SYSMIS;
+                     "MIN and MAX will be ignored."), pie->x_min, pie->x_max);
+          pie->x_min = pie->x_max = SYSMIS;
         }
     }
-
 }
 
 /* Add data from case C to the frequency table. */
@@ -464,26 +478,23 @@ postcalc (struct frq_proc *frq, const struct dataset *ds)
   for (i = 0; i < frq->n_vars; i++)
     {
       struct var_freqs *vf = &frq->vars[i];
-      int n_categories;
 
-      postprocess_freq_tab (vf);
+      postprocess_freq_tab (frq, vf);
 
       /* Frequencies tables. */
-      n_categories = vf->tab.n_valid + vf->tab.n_missing;
-      if  (cmd.table == FRQ_TABLE
-           || (cmd.table == FRQ_LIMIT && n_categories <= cmd.limit))
+      if (vf->tab.n_valid + vf->tab.n_missing <= frq->max_categories)
         dump_freq_table (vf, wv);
 
       /* Statistics. */
-      if (n_stats)
+      if (frq->n_stats)
 	dump_statistics (frq, vf, wv);
 
-      if (cmd.sbc_histogram && var_is_numeric (vf->var) && vf->tab.n_valid > 0)
+      if (frq->hist && var_is_numeric (vf->var) && vf->tab.n_valid > 0)
 	{
 	  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);
 
@@ -492,41 +503,19 @@ postcalc (struct frq_proc *frq, const struct dataset *ds)
                                vf->tab.valid_cases,
                                d[FRQ_MEAN],
                                d[FRQ_STDDEV],
-                               hist.draw_normal));
+                               frq->hist->draw_normal));
 
 	  statistic_destroy (&histogram->parent);
 	}
 
-      if (cmd.sbc_piechart)
-        do_piechart(vf->var, &vf->tab);
+      if (frq->pie)
+        do_piechart(frq->pie, vf->var, &vf->tab);
 
       cleanup_freq_tab (vf);
 
     }
 }
 
-/* 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
@@ -538,12 +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 (struct var_freqs *vf)
+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;
@@ -559,9 +575,12 @@ postprocess_freq_tab (struct var_freqs *vf)
   ft->n_missing = count - ft->n_valid;
 
   /* Sort data. */
-  compare = get_freq_comparator (cmd.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;
@@ -599,10 +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. */
@@ -643,8 +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 (;;)
       {
@@ -661,7 +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;
@@ -674,11 +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."));
@@ -718,14 +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;
@@ -773,122 +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. */
 
@@ -977,114 +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];
+          if (rank <= tp)
+            break;
 
-      /* Catches the case when p == 100% */
-      if ( ! pc->flag2 )
-	pc->x1 = pc->x2 = f->value.f;
-
-      /*
-      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;
@@ -1146,9 +1018,10 @@ 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, n_stats + frq->n_show_percentiles + 2);
+  t = tab_create (3, frq->n_stats + frq->n_show_percentiles + 2);
 
   tab_box (t, TAL_1, TAL_1, -1, -1 , 0 , 0 , 2, tab_nr(t) - 1) ;
 
@@ -1159,7 +1032,7 @@ dump_statistics (const struct frq_proc *frq, const struct var_freqs *vf,
   r=2; /* N missing and N valid are always dumped */
 
   for (i = 0; i < FRQ_N_STATS; i++)
-    if (stats & BIT_INDEX (i))
+    if (frq->stats & BIT_INDEX (i))
       {
 	tab_text (t, 0, r, TAB_LEFT | TAT_TITLE,
 		      gettext (st_name[i].s10));
@@ -1254,12 +1127,12 @@ freq_tab_to_hist (const struct frq_proc *frq, const struct freq_tab *ft,
   valid_freq = 0;
   for (i = 0; i < ft->n_valid; i++)
     {
-      const struct freq *frq = &ft->valid[i];
-      if (chart_includes_value (&hist, var, &frq->value))
+      const struct freq *f = &ft->valid[i];
+      if (chart_includes_value (frq->hist, var, &f->value))
         {
-          x_min = MIN (x_min, frq->value.f);
-          x_max = MAX (x_max, frq->value.f);
-          valid_freq += frq->count;
+          x_min = MIN (x_min, f->value.f);
+          x_max = MAX (x_max, f->value.f);
+          valid_freq += f->count;
         }
     }
 
@@ -1280,19 +1153,19 @@ freq_tab_to_hist (const struct frq_proc *frq, const struct freq_tab *ft,
   histogram = histogram_create (bins, x_min, x_max);
   for (i = 0; i < ft->n_valid; i++)
     {
-      const struct freq *frq = &ft->valid[i];
-      if (chart_includes_value (&hist, var, &frq->value))
-        histogram_add (histogram, frq->value.f, frq->count);
+      const struct freq *f = &ft->valid[i];
+      if (chart_includes_value (frq->hist, var, &f->value))
+        histogram_add (histogram, f->value.f, f->count);
     }
 
   return histogram;
 }
 
 static int
-add_slice (const struct freq *freq, const struct variable *var,
-           struct slice *slice)
+add_slice (const struct frq_chart *pie, const struct freq *freq,
+           const struct variable *var, struct slice *slice)
 {
-  if (chart_includes_value (&pie, var, &freq->value))
+  if (chart_includes_value (pie, var, &freq->value))
     {
       ds_init_empty (&slice->label);
       var_append_value_name (var, &freq->value, &slice->label);
@@ -1308,7 +1181,8 @@ add_slice (const struct freq *freq, const struct variable *var,
    The caller is responsible for freeing slices
 */
 static struct slice *
-freq_tab_to_slice_array(const struct freq_tab *frq_tab,
+freq_tab_to_slice_array(const struct frq_chart *pie,
+                        const struct freq_tab *frq_tab,
 			const struct variable *var,
 			int *n_slicesp)
 {
@@ -1320,9 +1194,9 @@ freq_tab_to_slice_array(const struct freq_tab *frq_tab,
   n_slices = 0;
 
   for (i = 0; i < frq_tab->n_valid; i++)
-    n_slices += add_slice (&frq_tab->valid[i], var, &slices[n_slices]);
+    n_slices += add_slice (pie, &frq_tab->valid[i], var, &slices[n_slices]);
   for (i = 0; i < frq_tab->n_missing; i++)
-    n_slices += add_slice (&frq_tab->missing[i], var, &slices[n_slices]);
+    n_slices += add_slice (pie, &frq_tab->missing[i], var, &slices[n_slices]);
 
   *n_slicesp = n_slices;
   return slices;
@@ -1332,12 +1206,13 @@ freq_tab_to_slice_array(const struct freq_tab *frq_tab,
 
 
 static void
-do_piechart(const struct variable *var, const struct freq_tab *frq_tab)
+do_piechart(const struct frq_chart *pie, const struct variable *var,
+            const struct freq_tab *frq_tab)
 {
   struct slice *slices;
   int n_slices, i;
 
-  slices = freq_tab_to_slice_array (frq_tab, var, &n_slices);
+  slices = freq_tab_to_slice_array (pie, frq_tab, var, &n_slices);
 
   if (n_slices < 2)
     msg (SW, _("Omitting pie chart for %s, which has only %d unique values."),