sort: Change priority queue implementation from heap to binary tree.

This is not so great in itself--it is probably a wash for performance,
and might be a loss--but it does make it possible to discover duplicate
sort keys while reading the data, which will be useful in an upcoming
commit.

diff --git a/src/math/sort.c b/src/math/sort.c
index 8719d877e096fcc9fc0d0f911d462d7bd0566725..77d9ea5cc190e42bed25d79a0f8c904e5874362b 100644 (file)
--- a/src/math/sort.c
+++ b/src/math/sort.c
@@ -1,5 +1,5 @@
 /* PSPP - a program for statistical analysis.
-   Copyright (C) 1997-9, 2000, 2006, 2009 Free Software Foundation, Inc.
+   Copyright (C) 1997-9, 2000, 2006, 2009, 2010 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
@@ -16,21 +16,21 @@
 
 #include <config.h>
 
-#include "sort.h"
+#include "math/sort.h"
 
 #include <stdio.h>
 
-#include <data/case.h>
-#include <data/casereader.h>
-#include <data/casewriter.h>
-#include <data/casewriter-provider.h>
-#include <data/settings.h>
-#include <data/subcase.h>
-#include <libpspp/array.h>
-#include <libpspp/assertion.h>
-#include <math/merge.h>
+#include "data/case.h"
+#include "data/casereader.h"
+#include "data/casewriter.h"
+#include "data/casewriter-provider.h"
+#include "data/settings.h"
+#include "data/subcase.h"
+#include "libpspp/assertion.h"
+#include "libpspp/bt.h"
+#include "math/merge.h"
 
-#include "xalloc.h"
+#include "gl/xalloc.h"
 
 #include "gettext.h"
 #define _(msgid) gettext (msgid)
@@ -197,21 +197,22 @@ sort_execute_1var (struct casereader *input, const struct variable *var)
 struct pqueue
   {
     struct subcase ordering;
-    struct pqueue_record *records;
-    size_t record_cnt;
+    struct bt bt;
     size_t record_cap;
     casenumber idx;
   };
 
 struct pqueue_record
   {
+    struct bt_node bt_node;
     casenumber id;
     struct ccase *c;
     casenumber idx;
   };
 
-static int compare_pqueue_records_minheap (const void *a, const void *b,
-                                           const void *pq_);
+static int compare_pqueue_records (const struct bt_node *a,
+                                   const struct bt_node *b,
+                                   const void *ordering);
 
 static struct pqueue *
 pqueue_create (const struct subcase *ordering, const struct caseproto *proto)
@@ -225,8 +226,7 @@ pqueue_create (const struct subcase *ordering, const struct caseproto *proto)
     pq->record_cap = max_buffers;
   else if (pq->record_cap < min_buffers)
     pq->record_cap = min_buffers;
-  pq->record_cnt = 0;
-  pq->records = xnmalloc (pq->record_cap, sizeof *pq->records);
+  bt_init (&pq->bt, compare_pqueue_records, &pq->ordering);
   pq->idx = 0;
 
   return pq;
@@ -244,7 +244,6 @@ pqueue_destroy (struct pqueue *pq)
           case_unref (c);
         }
       subcase_destroy (&pq->ordering);
-      free (pq->records);
       free (pq);
     }
 }
@@ -252,13 +251,13 @@ pqueue_destroy (struct pqueue *pq)
 static bool
 pqueue_is_full (const struct pqueue *pq)
 {
-  return pq->record_cnt >= pq->record_cap;
+  return bt_count (&pq->bt) >= pq->record_cap;
 }
 
 static bool
 pqueue_is_empty (const struct pqueue *pq)
 {
-  return pq->record_cnt == 0;
+  return bt_is_empty (&pq->bt);
 }
 
 static void
@@ -268,43 +267,42 @@ pqueue_push (struct pqueue *pq, struct ccase *c, casenumber id)
 
   assert (!pqueue_is_full (pq));
 
-  r = &pq->records[pq->record_cnt++];
+  r = xmalloc (sizeof *r);
   r->id = id;
   r->c = c;
   r->idx = pq->idx++;
-
-  push_heap (pq->records, pq->record_cnt, sizeof *pq->records,
-             compare_pqueue_records_minheap, pq);
+  bt_insert (&pq->bt, &r->bt_node);
 }
 
 static struct ccase *
 pqueue_pop (struct pqueue *pq, casenumber *id)
 {
   struct pqueue_record *r;
+  struct ccase *c;
 
   assert (!pqueue_is_empty (pq));
 
-  pop_heap (pq->records, pq->record_cnt--, sizeof *pq->records,
-            compare_pqueue_records_minheap, pq);
-
-  r = &pq->records[pq->record_cnt];
+  r = bt_data (bt_first (&pq->bt), struct pqueue_record, bt_node);
+  bt_delete (&pq->bt, &r->bt_node);
   *id = r->id;
-  return r->c;
+  c = r->c;
+  free (r);
+  return c;
 }
 
 /* Compares record-run tuples A and B on id, then on case data,
-   then on insertion order, in descending order. */
+   then on insertion order. */
 static int
-compare_pqueue_records_minheap (const void *a_, const void *b_,
-                                const void *pq_)
+compare_pqueue_records (const struct bt_node *a_, const struct bt_node *b_,
+                        const void *ordering_)
 {
-  const struct pqueue_record *a = a_;
-  const struct pqueue_record *b = b_;
-  const struct pqueue *pq = pq_;
+  const struct pqueue_record *a = bt_data (a_, struct pqueue_record, bt_node);
+  const struct pqueue_record *b = bt_data (b_, struct pqueue_record, bt_node);
+  const struct subcase *ordering = ordering_;
   int result = a->id < b->id ? -1 : a->id > b->id;
   if (result == 0)
-    result = subcase_compare_3way (&pq->ordering, a->c, &pq->ordering, b->c);
+    result = subcase_compare_3way (ordering, a->c, ordering, b->c);
   if (result == 0)
     result = a->idx < b->idx ? -1 : a->idx > b->idx;
-  return -result;
+  return result;
 }