SQL - ALTER DATABASE

The Alter Database command update database settings.

Syntax

ALTER DATABASE <attribute-name> <attribute-value>

Where: attribute-value attribute's value to set and attribute-name between those supported:

• STATUS database's status between:
• IMPORTING to set importing status
• DEFAULTCLUSTERID to set the default cluster. By default is 2 = "default"
• TYPE to change the database type. This makes only sense when you want to switch from any database to a graph specifying "graph"
• DATEFORMAT sets the default date format. Look at Java Date Format for more information. Default is "yyyy-MM-dd"
• DATETIMEFORMAT sets the default date time format. Look at Java Date Format for more information. Default is "yyyy-MM-dd HH:mm:ss"
• TIMEZONE set the default timezone. Look at Java Date TimeZones for more information. Default is the JVM's default timezone
• LOCALECOUNTRY sets the default locale country. Look at Java Locales for more information. Default is the JVM's default locale country. Example: "GB"
• LOCALELANGUAGE sets the default locale language. Look at Java Locales for more information. Default is the JVM's default locale language. Example: "en"
• CHARSET set the default charset charset. Look at Java Charset for more information. Default is the JVM's default charset. Example: "utf8"
• CLUSTERSELECTION sets the default strategy used on selecting the cluster where to create new records. This setting is read on class creation. After creation, each class has own modifiable strategy. Supported strategies are:
  • default, uses always the Class's defaultClusterId property. This was the default before 1.7
  • round-robin, put the Class's configured clusters in a ring and returns a different cluster every time restarting from the first when the ring is completed
  • balanced, checks the records in all the clusters and returns the smaller cluster. This allows the cluster to have all the underlying clusters balanced on size. On adding a new cluster to an existent class, the new empty cluster will be filled before the others because more empty then the others. In distributed configuration when configure clusters on different servers this setting allows to keep the server balanced with the same amount of data. Calculation of cluster size is made every 5 or more seconds to avoid to slow down insertion
• MINIMUMCLUSTERS, as the minimum number of clusters to create automatically when a new class is created. By default is 1, but on multi CPU/core having multiple clusters/files improves read/write performance
• CONFLICTSTRATEGY, (since 2.0) is the name of the strategy used to handle conflicts when OrientDB's MVCC finds an update or delete operation executed against an old record. The strategy is applied for the entire database, but single clusters can have own strategy (use ALTER CLUSTER command for this). While it's possible to inject custom logic by writing a Java class, the out of the box modes are:
  • version, the default, throw an exception when versions are different
  • content, in case the version is different checks if the content is changed, otherwise use the highest version and avoid throwing exception
  • automerge, merges the changes
• CUSTOM sets custom properties

  See also

• Console Command Create Database
• Console Command Drop Database

Examples

Change the database type to "graph"

ALTER DATABASE TYPE graph

Use GraphDB created with releases before 1.4

alter database custom useLightweightEdges=false
alter database custom useClassForEdgeLabel=false
alter database custom useClassForVertexLabel=false
alter database custom useVertexFieldsForEdgeLabels=false

History

1.7

• Added support for CLUSTERSELECTION that sets the strategy used on selecting the cluster where to create new records
• Added MINIMUMCLUSTERS to pre-create X clusters every time a new class is created