database/mariadb
1
0
Fork 0
mirror of https://github.com/MariaDB/server.git synced 2025-07-30 16:24:05 +03:00
Code Activity

better fix for MySQL bugs

Browse Source 
BUG#26447 prefer a clustered key for an index scan, as secondary index is always slower
  ... which was fixed to cause
BUG#35850 queries take 50% longer
  ... and was reverted

and

BUG#39653 prefer a secondary index for an index scan, as clustered key is always slower
  ... which was fixed to cause
BUG#55656 mysqldump takes six days instead of half an hour
  ... and was amended with a special case workaround


sql/opt_range.cc:
  move get_index_only_read_time() into the handler class
sql/sql_select.cc:
  use cost not an index length when choosing a cheaper index
This commit is contained in:
Sergei Golubchik
2010-10-20 12:58:43 +02:00
parent b56086cbbb
commit 60c15066db
4 changed files with 34 additions and 49 deletions
Download Patch File Download Diff File Expand all files Collapse all files

16
sql/handler.cc
View File

@ -2047,7 +2047,21 @@ handler *handler::clone(MEM_ROOT *mem_root)
return new_handler;
}
double handler::keyread_time(uint index, uint ranges, ha_rows rows)
{
/*
It is assumed that we will read trough the whole key range and that all
key blocks are half full (normally things are much better). It is also
assumed that each time we read the next key from the index, the handler
performs a random seek, thus the cost is proportional to the number of
blocks read. This model does not take into account clustered indexes -
engines that support that (e.g. InnoDB) may want to overwrite this method.
*/
double keys_per_block= (stats.block_size/2.0/
(table->key_info[index].key_length +
ref_length) + 1);
return (rows + keys_per_block - 1)/ keys_per_block;
}
void handler::ha_statistic_increment(ulong SSV::*offset) const
{