I had a customer recently who needed to reduce their database size on disk quickly without a lot of messy schema redesign and application recoding. They didn’t want to drop any actual data, and their index usage was fairly high, so we decided to look for unused indexes that could be removed. Collecting data It’s [...]
A case for MariaDB’s Hash Joins
May 31, 2012 By Leave a Comment
MariaDB 5.3/5.5 has introduced a new join type “Hash Joins” which is an implementation of a Classic Block-based Hash Join Algorithm. In this post we will see what the Hash Join is, how it works and for what types of queries would it be the right choice. I will show the results of executing benchmarks [...]
The case for getting rid of duplicate “sets”
May 17, 2011 By 2 Comments
The most useful feature of the relational database is that it allows us to easily process data in sets, which can be much faster than processing it serially. When the relational database was first implemented, write-ahead-logging and other technologies did not exist. This made it difficult to implement the database in a way that matched [...]
Checking the subset sum set problem with set processing
May 16, 2011 By Leave a Comment
Hi, Here is an easy way to run the subset sum check from SQL, which you can then distribute with Shard-Query:
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 | CREATE TABLE `the list` ( `id` bigint(20) NOT NULL AUTO_INCREMENT, `val` bigint(20) NOT NULL DEFAULT '0', PRIMARY KEY (`id`), KEY `id` (`id`) ) ENGINE=MyISAM; SELECT val as `val`, COUNT(DISTINCT (id)) as `cd` FROM test.data as d WHERE val in (-2,-3,-10,15,15,16) GROUP BY val; +-----+----------+----------+ | val | cd | CNT | +-----+----------+----------+ | -10 | 1 | 1 | | -3 | 1 | 1 | | -2 | 1 | 1 | | 15 | 35417088 | 35417088 | +-----+----------+----------+ 5 rows in set (40.20 sec) |
Notice there is no 16 in the list. We did not pass the check. There are enough 15s though. The distinct value count for each item in the output set, must at least [...]
Flexviews – part 3 – improving query performance using materialized views
April 4, 2011 By 20 Comments
Combating “data drift” In my first post in this series, I described materialized views (MVs). An MV is essentially a cached result set at one point in time. The contents of the MV will become incorrect (out of sync) when the underlying data changes. This loss of synchronization is sometimes called drift. This is conceptually [...]
Tuning InnoDB Concurrency Tickets
May 24, 2010 By 5 Comments
InnoDB has an oft-unused parameter innodb_concurrency_tickets that seems widely misunderstood. From the docs: “The number of threads that can enter InnoDB concurrently is determined by the innodb_thread_concurrency variable. A thread is placed in a queue when it tries to enter InnoDB if the number of threads has already reached the concurrency limit. When a thread [...]
A workaround for the performance problems of TEMPTABLE views
May 19, 2010 By 7 Comments
MySQL supports two different algorithms for views: the MERGE algorithm and the TEMPTABLE algorithm. These two algorithms differ greatly. A view which uses the MERGE algorithm can merge filter conditions into the view query itself. This has significant performance advantages over TEMPTABLE views. A view which uses the TEMPTABLE algorithm will have to compute the [...]
InnoDB: look after fragmentation
November 5, 2009 By 12 Comments
One problem made me puzzled for couple hours, but it was really interesting to figure out what’s going on. So let me introduce problem at first. The table is
1 2 3 4 5 6 7 8 | CREATE TABLE `c` ( `tracker_id` int(10) unsigned NOT NULL, `username` char(20) character set latin1 collate latin1_bin NOT NULL, `time_id` date NOT NULL, `block_id` int(10) unsigned default NULL, PRIMARY KEY (`tracker_id`,`username`,`time_id`), KEY `block_id` (`block_id`) ) ENGINE=InnoDB |
Table has 11864696 rows and takes Data_length: 698,351,616 bytes on disk The problem is that after restoring table from mysqldump, the query that scans data [...]
Air traffic queries in InfiniDB: early alpha
November 2, 2009 By 19 Comments
As Calpont announced availability of InfiniDB I surely couldn’t miss a chance to compare it with previously tested databases in the same environment. See my previous posts on this topic: Analyzing air traffic performance with InfoBright and MonetDB Air traffic queries in LucidDB I could not run all queries against InfiniDB and I met some [...]
How (not) to find unused indexes
October 16, 2009 By 2 Comments
I’ve seen a few people link to an INFORMATION_SCHEMA query to be able to find any indexes that have low cardinality, in an effort to find out what indexes should be removed. This method is flawed – here’s the first reason why:
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 | CREATE TABLE `sales` ( `id` int(11) NOT NULL AUTO_INCREMENT, `customer_id` int(11) DEFAULT NULL, `status` enum('archived','active') DEFAULT NULL, PRIMARY KEY (`id`), KEY `status` (`status`) ) ENGINE=MyISAM AUTO_INCREMENT=65691 DEFAULT CHARSET=latin1; mysql> SELECT count(*), status FROM sales GROUP by status; +----------+---------+ | count(*) | status | +----------+---------+ |   65536 | archived | |     154 | active | +----------+---------+ 2 rows in set (0.17 sec) mysql> EXPLAIN SELECT * FROM sales WHERE status='active'; # query 1 +----+-------------+-------+------+---------------+--------+---------+-------+------+-------------+ | id | select_type | table | type | possible_keys | key   | key_len | ref  | rows | Extra      | +----+-------------+-------+------+---------------+--------+---------+-------+------+-------------+ | 1 | SIMPLE     | sales | ref | status       | status | 2      | const | 196 | Using where | +----+-------------+-------+------+---------------+--------+---------+-------+------+-------------+ 1 row in set (0.06 sec) mysql> EXPLAIN SELECT * FROM sales WHERE status='archived'; # query 2 +----+-------------+-------+------+---------------+------+---------+------+-------+-------------+ | id | select_type | table | type | possible_keys | key | key_len | ref | rows | Extra      | +----+-------------+-------+------+---------------+------+---------+------+-------+-------------+ | 1 | SIMPLE     | sales | ALL | status       | NULL | NULL   | NULL | 65690 | Using where | +----+-------------+-------+------+---------------+------+---------+------+-------+-------------+ 1 row in set (0.01 sec) |
The cardinality of status index is woeful, but provided that the application [...]
