You have to run the default task ('archives') of the build.xml provided
in the Speedo distribution.
By default at runtime, the first time Speedo is asked to manage a persistent class, it tries to initialize the data structure (SQL tables) associated to the class.
Nevertheless when using Speedo in an application server, some databases do not support the creation of the data structure inside a XA transaction. In this case, the user must create the data structure before the launching the server. To do this, Speedo provides a class which initializes the the data structure of persistent classes, outside the scope of a transaction. This class is named 'org.objectweb.speedo.tools.DataStructureCreation'. The Speedo examples show usages of this class through the ant target 'createDataStruct' provided with their build.xml. In fact, the main method of the class expects, as arguments, the names of classes to initialize (white space as separator).
java org.objectweb.speedo.tools.DataStructureCreation com.foo.MyPersistentClass1 com.foo.MyPersistentClass2
The speedo project uses the version 1.6 of Jarkata Ant. In addition you have to put the junit.jar (in the 'externals' directory) with the ant libraries ($ANT_HOME/lib).
Target | Description |
compile | compile the project in the ./output/build directory |
dist | create the speedo distribution in the ./output/dist directory |
zip | create the zip file corresponding to the project and the distribution. The zip file are created into the ./output/zip directory |
test | run all speedo tests |
Set the properties 'org.objectweb.speedo.query.prefetch.query' or
'org.objectweb.speedo.query.prefetch.extent' to 'off' in the
speedo.properties configuration file.
Some database schemas contains definition of integrity constraints like a foreign key definition. As other persistence containers, Speedo cannot support that the integrity constraint have to be check at each database action. You have to defer the chekcing at commit time. Indeed the persistence containers cannot known the right order to flush modification on the database. A following way shows a declaration of a foreign key constraint which is checked at commit time.
#declaration of a first table TA create table TA ( A_ID NUMBER(30) not null, constraint PK_MESSAGE primary key (A_ID)); # declaration of a first table TB create table TB ( B_ID NUMBER(30) not null, AID NUMBER(30), constraint PK_MESSAGE primary key (B_ID)); # declaration of a foreign key contraint with the checking deferred at commit time alter table TB add constraint MY_CONSTRAINT_NAME foreign key (AID) references TA (A_ID) initially deferred deferrable;
The important thing to see is the last point: initially deferred
deferrable, because it is this option which specifies that the
constraint is checked at commit time.
In some cases you want to have a Map of persistent classes indexed by the a field of the referenced classes such as the following example:
class A { /** * value = a B * key = B.f1 */ Map mapOfB; ... } class B { String f1 ... }
Speedo supports this mapping only in case of the B class has no sub
classes. To knwon how to specify this mapping see the section about
Map field in the user manual.
In some cases you want to map two persistent fields on the same column. Speedo permits this mapping except for two primitive fields. The way to declare two fields on the same column is very simple. You only have to specify the same column name for both fields
The following example shows the mapping of a primitive fields and a reference field on the same SQL column.
class A { String myid; String idb; B myb; ... } class B { String myid ... } #the .jdo file <class name="A"/> <field name="myid" primary-key="true"> <extension vendor-name="speedo" key="sql-name" value="PKA"/> </field> <field name="idb"> <!-- the field idb is mapped over the column FKB --> <extension vendor-name="speedo" key="sql-name" value="FKB"/> </field> <field name="myb"> <!-- the field myb is mapped over the column FKB which is a foreign key on the PKB column, and corresponds also to the idb field--> <extension vendor-name="speedo" key="target-foreign-keys" value="PKB=FKB"/> </field> </class> <class name="B"/> <field name="myid" primary-key="true"> <extension vendor-name="speedo" key="sql-name" value="PKB"/> </field> </class>
Note: The primitive field can compose the identifier of the current class.
class A { String ida1; String ida2; B myb; ... } class B { String myid ... } #the .jdo file <class name="A"/> <field name="ida1" primary-key="true"> <!-- the field ida1 is mapped over the column PKA1 --> <extension vendor-name="speedo" key="sql-name" value="PKA1"/> </field> <field name="ida2" primary-key="true"> <!-- the field ida2 is mapped over the column PKA2 --> <extension vendor-name="speedo" key="sql-name" value="PKA2"/> </field> <field name="myb"> <!-- the field myb is mapped over the column PKA2 which is a foreign key on the PKB column, and corresponds also to the ida2 field --> <extension vendor-name="speedo" key="target-foreign-keys" value="PKB=PKA2"/> </field> </class> <class name="B"/> <field name="myid" primary-key="true"> <extension vendor-name="speedo" key="sql-name" value="PKB"/> </field> </class>
Note: The possible couples are the following:
Speedo supports several relational databases (supported databases).
Speedo is based on JORM, and uses in particular RdbAdapter for each
supported database. A RdbAdapter permits to take in account non conformance
of JDBC driver/database to the JDBC/SQL standards. Then if your database is
not already supported you can easily write a RdbAdapter. An RdbAdapter is
a short class about 25 java lines. JORM provides a RdbAdaper tester
permiting to find the conformance problem, to solve and to verify that your
adapter works fine. Your are welcomed to contribute to the adapter suite
for adding new database support.
Download the RdbAdapter developpement kit
To store your persistent classes in other support type, you have to write a
JORM Mapper. Contact the JORM team for
more information.
Yes, Speedo supports clustering and concurrent database accesses (read or write). For more details read the Concurrency Management chapter in the usermanual.
Read the Caching chapter in the usermanual.