About MongoDB Schema-free vs MySQL DDL

The document database (Document-orient db) supports Schema-free, and his one-line record format is: {"name": "Design", "value": "Straight"}}, is actually using key/ The MONGO of value storage uses Bson binary encoding, so space should be more than JSON, parsing in JSON. Each document contains metadata information, which is not forced to use the same format between documents, and they also support a variety of indexes, at a convenient level that is more triumphant than MySQL. On the MySQL table design and MongoDB scheme free design comparison, can refer to the following: http://blog.nosqlfan.com/html/2985.html however, due to the large number of metadata redundancy, space is very serious waste.

For MySQL, the typical relational database, the impact of DDL on Io can not be overlooked, which is often caused by the data storage format.

The Innodb storage engine is already trying to modify the level two index, which means that data cannot be copied during the DDL process, but the table is locked during this period. This is handy for using the online schema change OSC.

The common practice is to do DDL operations on the repository now, and then restore the original after the master-slave switch, and so on.

Is that the relational storage engine stores metadata and data separately, and some modifications to the metadata require that all data be modified. One way to avoid this problem is to put metadata and data together, which is the popular schema-free in the NoSQL field.

This column more highlights: http://www.bianceng.cn/database/MySQL/