June 19, 2013

Post: InnoDB Full-text Search in MySQL 5.6: Part 2, The Queries!

2, The Queries![/caption] This is part 2… The same query, run against InnoDB on… 1. row *************************** Table…column the PK, but neither of those worked. Also, trying to set innodb_ft_user_stopword_table produced the samecompare

Post: Using CHAR keys for joins, how much is the overhead ?

….t1.i rows: 26 Extra: 2 rows in set (0.00 sec) So what about Innodb ? Innodb executed the same query in 2.9 seconds… it has to read row data from OS Cache. So what if we convert i and j columns to varchar while sticking… as default ? Joining on Char columns completes in 4.5 seconds on Innodb which is about 50% slower compared to Joining on Int, for MyISAM…

Post: Multi Column indexes vs Index Merge

on selectivity and correlation. If where clause from first column selects 5% of the rows and applying where clause on second columncompared to single index only but it is by far better to use multi column

Post: Improved InnoDB fast index creation

column to trigger table rebuilds without affecting the table size. mysql> SET expand_fast_index_creation=ON; Query OK, 0 rowssame time was spent oncompared … optimize | status | OK | +——–+———-+———-+——————————————————————-+ 2 rows in set (2

Post: How much overhead is caused by on disk temporary tables

… do on disk temporary tables cause compared to MEMORY tables. To have things comparable I … 5 rows in set (0.00 sec) The table contains 1.000.000 rows with mostly unique “c” column…. 0 0 5 44 96 103 2 0 94 3 0 0 3 … the same queries: mysql> set global key_buffer_size=8000000; Query OK, 0 rows affected…

Post: The story of one MySQL Upgrade

ON DUPLICATE KEY UPDATE query instances, and only one of them into table with AUTO_INCREMENT column… replication to check consistency but comparing 2 servers directly is faster … stopped both slaves at the same position again and used … check result sets. –compare-results-method rows helped to remove …

Post: EXPLAIN EXTENDED can tell you all kinds of interesting things

column we see that MySQL mentions ‘const’ tables. A ‘const’ table is a table that contains 0 or 1 rows, or a table onrow. The contents of these tables are read before the query begins executing. Because of this, MySQL can comparesame

Post: A case for MariaDB's Hash Joins

… for each row scanned performing a lookup in the hash table by using values of the columns participating …and the OS installed is Centos 6.2 Table Structure Before moving on, let’s take a look at … compared to doing batched index lookups (BKA). However, in Test C the cost of hash join remains the same

Post: Join Optimizations in MySQL 5.6 and MariaDB 5.5

… the same page multiple times. Consider if you have a small buffer pool and a very large number of rows … 13384 Innodb_data_read 5G 5.4G 2.2G 5.4G 2.2G 2.2G Innodb_data_reads 329115 355323 143808 … query on MariaDB 5.5 is slow as compared to MySQL 5.6. Next interesting thing are the last two columns

Post: InnoDB Full-text Search in MySQL 5.6 (part 1)

compare query results between MyISAM FTS and InnoDB FTS over the samecolumn called FTS_DOC_ID which is a BIGINT UNSIGNED NOT NULL with a unique index on…dir_test_innodb | optimize | status | OK | +———————-+———-+———-+———-+ 1 row in set (4.60 …. In part 2 of this series…