Fix for bug#5138: hash indexes on heap tables support statistics.

 KEY::rec_per_key is updated every time 1/HEAP_STATS_UPDATE_THRESHOLD part of table records has been changed.

3 files changed, 64 insertions, 9 deletions

diff --git a/sql/ha_heap.cc b/sql/ha_heap.cc
index d7327362286..7643037e24f 100644
--- a/sql/ha_heap.cc
+++ b/sql/ha_heap.cc
@@ -30,6 +30,12 @@
 const char **ha_heap::bas_ext() const
 { static const char *ext[1]= { NullS }; return ext; }
 
+/*
+  Hash index statistics is updated (copied from HP_KEYDEF::hash_buckets to 
+  rec_per_key) after 1/HEAP_STATS_UPDATE_THRESHOLD records have been inserted/
+  updated/deleted. delete_all_rows() and table flush cause immediate update.
+*/
+#define HEAP_STATS_UPDATE_THRESHOLD 10
 
 int ha_heap::open(const char *name, int mode, uint test_if_locked)
 {
@@ -48,6 +54,8 @@ int ha_heap::open(const char *name, int mode, uint test_if_locked)
   {
     /* Initialize variables for the opened table */
     set_keys_for_scanning();
+    if (table->tmp_table == NO_TMP_TABLE)
+      update_key_stats();
   }
   return (file ? 0 : 1);
 }
@@ -84,28 +92,59 @@ void ha_heap::set_keys_for_scanning(void)
   }
 }
 
+void ha_heap::update_key_stats()
+{
+  printf("update_key_stats\n");
+  for (uint i= 0; i < table->keys; i++)
+  {
+    KEY *key=table->key_info+i;
+    if (key->algorithm != HA_KEY_ALG_BTREE)
+    {
+      ha_rows hash_buckets= file->s->keydef[i].hash_buckets;
+      key->rec_per_key[key->key_parts-1]= 
+        hash_buckets ? file->s->records/hash_buckets : 0;
+    }
+  }
+  records_changed= 0;
+}
+
 int ha_heap::write_row(byte * buf)
 {
+  int res;
   statistic_increment(ha_write_count,&LOCK_status);
   if (table->timestamp_default_now)
     update_timestamp(buf+table->timestamp_default_now-1);
   if (table->next_number_field && buf == table->record[0])
     update_auto_increment();
-  return heap_write(file,buf);
+  res= heap_write(file,buf);
+  if (!res && table->tmp_table == NO_TMP_TABLE && 
+      ++records_changed*HEAP_STATS_UPDATE_THRESHOLD > file->s->records)
+    update_key_stats();
+  return res;
 }
 
 int ha_heap::update_row(const byte * old_data, byte * new_data)
 {
+  int res;
   statistic_increment(ha_update_count,&LOCK_status);
   if (table->timestamp_on_update_now)
     update_timestamp(new_data+table->timestamp_on_update_now-1);
-  return heap_update(file,old_data,new_data);
+  res= heap_update(file,old_data,new_data);
+  if (!res && table->tmp_table == NO_TMP_TABLE && 
+      ++records_changed*HEAP_STATS_UPDATE_THRESHOLD > file->s->records)
+    update_key_stats();
+  return res;
 }
 
 int ha_heap::delete_row(const byte * buf)
 {
+  int res;
   statistic_increment(ha_delete_count,&LOCK_status);
-  return heap_delete(file,buf);
+  res= heap_delete(file,buf);
+  if (!res && table->tmp_table == NO_TMP_TABLE && 
+      ++records_changed*HEAP_STATS_UPDATE_THRESHOLD > file->s->records)
+    update_key_stats();
+  return res;
 }
 
 int ha_heap::index_read(byte * buf, const byte * key, uint key_len,
@@ -227,6 +266,8 @@ int ha_heap::extra(enum ha_extra_function operation)
 int ha_heap::delete_all_rows()
 {
   heap_clear(file);
+  if (table->tmp_table == NO_TMP_TABLE)
+    update_key_stats();
   return 0;
 }
 
@@ -383,7 +424,14 @@ ha_rows ha_heap::records_in_range(uint inx, key_range *min_key,
       min_key->flag != HA_READ_KEY_EXACT ||
       max_key->flag != HA_READ_AFTER_KEY)
     return HA_POS_ERROR;			// Can only use exact keys
-  return 10;					// Good guess
+  else
+  {
+    ha_rows records= file->s->records;
+    if (!records)
+      return 0;
+    ha_rows res= records / file->s->keydef[inx].hash_buckets;
+    return res ? res : 1;
+  }
 }
 
 
diff --git a/sql/ha_heap.h b/sql/ha_heap.h
index 9ca6b9b76b6..f36e9f31c55 100644
--- a/sql/ha_heap.h
+++ b/sql/ha_heap.h
@@ -27,9 +27,10 @@ class ha_heap: public handler
 {
   HP_INFO *file;
   key_map btree_keys;
-
- public:
-  ha_heap(TABLE *table): handler(table), file(0) {}
+  /* number of records changed since last statistics update */
+  uint    records_changed;
+public:
+  ha_heap(TABLE *table): handler(table), file(0), records_changed(0) {}
   ~ha_heap() {}
   const char *table_type() const { return "HEAP"; }
   const char *index_type(uint inx)
@@ -91,5 +92,6 @@ class ha_heap: public handler
 
   THR_LOCK_DATA **store_lock(THD *thd, THR_LOCK_DATA **to,
 			     enum thr_lock_type lock_type);
-
+private:
+  void update_key_stats();
 };
diff --git a/sql/structs.h b/sql/structs.h
index c30d85f59cb..846b3400fab 100644
--- a/sql/structs.h
+++ b/sql/structs.h
@@ -89,7 +89,12 @@ typedef struct st_key {
   enum  ha_key_alg algorithm;
   KEY_PART_INFO *key_part;
   char	*name;				/* Name of key */
-  ulong *rec_per_key;			/* Key part distribution */
+  /*
+    Array of AVG(#records with the same field value) for 1st ... Nth key part.
+    0 means 'not known'.
+    For temporary heap tables this member is NULL.
+  */
+  ulong *rec_per_key;
   union {
     int  bdb_return_if_eq;
   } handler;