An abstract SQL type to be used when automatically creating a table for the entity bean. Resin-CMP will use the abstract-sql-type and the database driver type to determine the actual type used to create the database table.

Defines the size of the entity cache.

The default cache-size is 8192.

Defines how long a loaded bean will remain loaded. As long as a cached CMP bean is loaded, it doesn't need to query the database, saving time and database load.

The default value is 5 seconds.

Defines one of the bean's fields to be managed by Resin-CMP. Since Resin-CMP will introspect the cmp-fields from the classes, the cmp-field is optional.

cmp-fields are either Java primitives or serializable types, i.e. values in a database table. Relations between beans are configured in the relations section.

Element	Meaning
field-name	the field name
sql-column	the SQL column name
sql-type	the SQL type

The SQL name is generated automatically from the cmp-field (or the getter method.) The getter method getMyField corresponds to the cmp-field myField and the SQL column my_field. An EJB-QL query, always uses the cmp-field name, e.g. myField.

The cmp-field can specify the SQL column name using sql-column.

student.ejb

<entity> <abstract-schema-name>students</abstract-schema-name> <sql-table>student_table</sql-table> ... <cmp-field> <field-name>id</field-name> <sql-column>student_id</sql-column> </cmp-field> <cmp-field> <field-name>name</field-name> <sql-column>user_name</sql-column> </cmp-field> ... </entity>

The above descriptor would expect the following schema:

student.sql

CREATE TABLE student_table ( student_id INTEGER AUTO_INCREMENT, user_name VARCHAR(255), PRIMARY KEY(student_id) );

Specifies the database JNDI name for the DataSource. If unspecified, it defaults to the value in EJBServer.

Specifies the field name for a cmp-field definition. The field-name is the bean name for the Java getter method. So the field-name for getFoo is foo.

Configuration for a business method. Each business method is specified by a <signature>. The signature syntax is the same as Java function syntax without the return type.

The arguments are optional and the method name can be "*". The best matching method will be used. Methods which match a method with specified args is better than a method with only a method name. And both are better than the default match.

Configure most methods as read-only

<method signature="*"> <resin-isolation>read-only</resin-isolation> </method> <method signature="writeMethod(String)"> <resin-isolation>database</resin-isolation> </method>

method ::= signature, resin-isolation?

Adds a new function to the EJB-QL language. The query-function must specify the function signature in Java syntax. The following types are allowed: int, double, boolean, String, Date and any.

An optional sql attribute lets you provide custom functions.

Adding MD5

<query-function> <signature>String MD5(String)</signature> </query-function>

Adding an add function

<query-function> <signature>int add(int, int)</signature> <sql>((?1) + (?2))</signature> </query-function>

query-function = element query-function { signature, sql? }

If true, an ejbSelect or find query will load the beans in the query's select. If false, the query will only load the primary key and lazily load the bean when it's referenced.

Normally, setting this true will give better performance.

Default all querys to lazy load

<method> <signature>*</signature> <query-loads-bean>false</query-loads-bean> </method>

Default: true by default.

Configures a bean as read-only. Read-only entity bean values can be cached between calls even in "update" transactions . The cache time is configured with cache-time.

Top-level configuration for the Resin-specific configuration.

resin-ejb ::= enterprise-beans

Configures the isolation level for a method. Each method is classified with either read-only or database isolation. read-only methods only read database from the database. database methods read and write to the database.

By telling Resin which methods are read-only, Resin can effectively cache read values and reduce the database load.

For methods defined by EJB as write methods, Resin will ignore the read-only value. setXXX, create, and remove methods always have database or stronger isolation.

For methods defined by EJB as read methods, Resin will use read-only as a default.

Default: the default value is database

Speficies a function signature, i.e. a method declaration, in Java syntax. In the case of <query-function>, <signature> specifies the name and types for a new EJB-QL function.

The SQL column name for a field or relation. If is unspecified, it defaults to the field name. The value of sql-column is never used in the EJB queries. It's just used during SQL generation.

For relations where the target bean has a compound primary key, the sql-column can specify a references attribute to distinguish which part of the primary key to configure.

Specifying a cmp-field

<entity> <ejb-name>students</ejb-name> ... <cmp-field> <field-name>id</field-name> <sql-column>student_id</sql-column> </cmp-field> </entity>

A pointer to a compound-key bean

<cmr-field> <cmr-field-name>child</cmr-field-name> <sql-column references='parent'>test_parent</sql-column> <sql-column references='name'>test_cname</sql-column> </cmr-field>

The SQL type to be used when automatically creating a table for the entity bean. The sql-type is the string that will be used in the CREATE TABLE.