MDEV-36761 Implement NULL-aware cardinality estimation for indexed columns

(Preparation for the main patch)

set_statistics_for_table() incorrectly treated indexes with all NULL
values the same as indexes with no statistics, because avg_frequency
is 0 in both cases. This caused the optimizer to ignore valid EITS
data and fall back to engine statistics.

Additionally, KEY::actual_rec_per_key() would fall back to engine
statistics even when EITS was available, and used incorrect pointer
comparison (rec_per_key == 0 instead of nullptr).

Fix by adding Index_statistics::stats_were_read flag to track per-index
whether statistics were actually read from persistent tables, and
restructuring actual_rec_per_key() to prioritize EITS when available.

Author: mariadb-OlegSmirnov

sql/sql_statistics.cc

@@ -3120,7 +3120,10 @@ read_statistics_for_table(THD *thd, TABLE *table,
       found|= index_stat.get_stat_values(index_statistics);
     }
     if (found)
+    {
       new_stats_cb->stats_available|= TABLE_STAT_INDEX;
+      index_statistics->mark_stats_as_read();
+    }
 
     key_part_map ext_key_part_map= key_info->ext_key_part_map;
     if (key_info->user_defined_key_parts != key_info->ext_key_parts &&
@@ -4157,7 +4160,7 @@ void set_statistics_for_table(THD *thd, TABLE *table)
       (check_eits_preferred(thd) &&
        table->stats_is_read &&
        key_info->read_stats->avg_frequency_is_inited() &&
-       key_info->read_stats->get_avg_frequency(0) > 0.5);
+       key_info->read_stats->has_stats());
   }
 }
 

sql/sql_statistics.h

@@ -603,10 +603,19 @@ class Index_statistics
     k-component prefixes among them 
   */
   ulonglong *avg_frequency;
+  bool stats_were_read;
 
 public:
 
-  void init_avg_frequency(ulonglong *ptr) { avg_frequency= ptr; }
+  void init_avg_frequency(ulonglong *ptr)
+  {
+    avg_frequency= ptr;
+    stats_were_read= false;
+  }
+
+  void mark_stats_as_read() { stats_were_read= true; }
+
+  bool has_stats() const { return stats_were_read; }
 
   bool avg_frequency_is_inited() { return avg_frequency != NULL; }
 

sql/table.cc

@@ -10299,10 +10299,15 @@ uint TABLE_SHARE::actual_n_key_parts(THD *thd)
 
 double KEY::actual_rec_per_key(uint i) const
 { 
-  if (rec_per_key == 0)
-    return 0;
-  return (is_statistics_from_stat_tables ?
-          read_stats->get_avg_frequency(i) : (double) rec_per_key[i]);
+  if (is_statistics_from_stat_tables)
+  {
+    // Use engine-independent statistics (EITS)
+    return read_stats->get_avg_frequency(i);
+  }
+  // Fall back to engine-dependent statistics if EITS is not available
+  if (rec_per_key == nullptr)
+    return 0; // No statistics available
+  return (double) rec_per_key[i];
 }
 
 /*