Detailed explanation of storage engines InnoDB and MyISAM in MySQL

1.InnoDB storage engine

InnoDB becomes the default storage engine in MySQL5.5. InnoDB is a transactional storage engine designed to handle a large number of short-term transactions. Its performance and automatic crash recovery features make it popular in non-transactional scenarios as well.

So you can say: "Unless you need to use some features that InnoDB does not have, and there is no other way to replace it, you should give priority to it".

InnoDB data is stored in the table space, which can store the data and indexes of each table in separate files.

InnoDB uses MVCC to support high concurrency and implements four standard isolation levels. The default level is REPEATABLE READ (repeatable read), and the gap lock strategy prevents phantom reads.

InnoDB is built based on a clustered index. The clustered index has high performance for primary key queries.

InnoDB supports true hot backup, which can be achieved through the XtraBackup tool.

InnoDB has made many internal optimizations, including predictable read-ahead when reading data from disk, the ability to automatically create hash indexes in memory, and insertion buffers that can speed up insertion operations.

Application scenario:

                  Requires online hot backup

　　 The amount of data is large and rapid recovery is required after the system crashes. Such as order processing.

2.MyISAM Storage Engine

Before MySQL5.1, MyISAM was the default storage engine. MyISAM does not support transactions and row-level locks, and its biggest flaw is that it cannot safely recover after a crash.

.MyISAM stores tables in two files: data files and index files.

MyISAM locks the entire table, adds shared locks to all required tables when reading, and adds exclusive locks when writing. However, when there is a read query on the table, you can also insert new records into the table (concurrent insertion)

MyISAM can perform check and repair operations manually or automatically, but it may cause data loss and the repair operation is very slow.

Long fields such as BLOB and TEXT in the MyISAM table can also create indexes based on the first 500 characters. MyISAM also supports full-text indexing, which is an index created based on word segmentation and can support complex queries.

If MyISAM specifies DELAY_KEY_WRITE (delayed update of index key) when creating a table, the index will be written to the key buffer in memory when each modification is completed, blocking the key buffer from being cleared or the table being closed. Only then will the corresponding index block be written to the disk. This method can greatly improve writing performance. But indexes can become corrupted when the database or host crashes.

Application scenarios:

Mainly scenarios for SELECT and INSERT operations, such as general log-type applications.

       

The above is the detailed content of Detailed explanation of storage engines InnoDB and MyISAM in MySQL. For more information, please follow other related articles on the PHP Chinese website!