database/mariadb
1
0
Fork 0
mirror of https://github.com/MariaDB/server.git synced 2025-08-08 11:22:35 +03:00
Code Activity

MDEV-11172: EXPLAIN shows non-sensical value for key_len with type=index

Browse Source 
The issue happens when the secondary keys are extended with primary
key parts. Inside the function TABLE_SHARE::init_from_binary_frm_image()
adds the length bytes for the primary key key parts to the length of the
secondary key. This is not needed because when the extended keys are
used we recalculate the length for the used key parts.

Also removed TABLE_SHARE::total_key_length as it is not used in the code

Apporved-by: Monty <monty@mariadb.org>
This commit is contained in:
Varun Gupta
2021-01-29 00:27:19 +05:30
parent 0921656734
commit b87c342da5
5 changed files with 45 additions and 4 deletions
Download Patch File Download Diff File Expand all files Collapse all files

10
sql/table.cc
View File

@@ -2273,6 +2273,12 @@ int TABLE_SHARE::init_from_binary_frm_image(THD *thd, bool write,
key_part->null_bit= field->null_bit;
key_part->store_length+=HA_KEY_NULL_LENGTH;
keyinfo->flags|=HA_NULL_PART_KEY;
/*
This branch is executed only for user defined key parts of the
secondary indexes.
*/
DBUG_ASSERT(i < keyinfo->user_defined_key_parts);
keyinfo->key_length+= HA_KEY_NULL_LENGTH;
}
if (field->type() == MYSQL_TYPE_BLOB ||
@@ -2285,7 +2291,8 @@ int TABLE_SHARE::init_from_binary_frm_image(THD *thd, bool write,
else
key_part->key_part_flag|= HA_VAR_LENGTH_PART;
key_part->store_length+=HA_KEY_BLOB_LENGTH;
keyinfo->key_length+= HA_KEY_BLOB_LENGTH;
if (i < keyinfo->user_defined_key_parts)
keyinfo->key_length+= HA_KEY_BLOB_LENGTH;
}
if (field->type() == MYSQL_TYPE_BIT)
key_part->key_part_flag|= HA_BIT_PART;
@@ -2378,7 +2385,6 @@ int TABLE_SHARE::init_from_binary_frm_image(THD *thd, bool write,
set_if_bigger(share->max_key_length,keyinfo->key_length+
keyinfo->user_defined_key_parts);
share->total_key_length+= keyinfo->key_length;
/*
MERGE tables do not have unique indexes. But every key could be
an unique index on the underlying MyISAM table. (Bug #10400)