RDB DAS C++ - ConventionOverConfiguration (TUSCANY) created by Adriano Crestani
http://cwiki.apache.org/confluence/pages/viewpage.action?pageId=69443
Content:
---------------------------------------------------------------------
h1. ConventionOverConfiguration
This concept reflects the notion that configuration (defining a bunch of stuff
in a side file) can be avoided by following convention (adhering to prescribed
patterns) and that this is generally a good thing. The RDB DAS has several
reasonable conventions in place and using these can significantly reduce a
developer's work-load.
As an example, the RDB DAS provides a straight-forward, one-to-one mapping of
database tables to DataObject Types and from table columns to DataObject
properties. If a user provides no mapping information to the contrary, a graph
of DataObjects returned by the DAS will directly map Types/Properties to the
Tables/Columns queried.
For example, consider a database with the following table:
{code}
TABLE CUSTOMER
ID INTEGER
LASTNAME VARCHAR(20)
ADDRESS VARCHAR(30)
{code}
If an application uses the DAS to query this table with the following SQL
statement:
{noformat}
SELECT * FROM CUSTOMER
{noformat}
The DAS will return a graph of DataObjects each of Type CUSTOMER and each
instance in will have three properties: (ID, LASTNAME, ADDRESS). Here is the
example code. Note that no configuration data is provided at all.
{panel}
DASImpl\* das = (DASImpl*) DASImpl::getFactory().createDAS(getConnection());
CommandPtr readCustomers = das->createCommand("select * from CUSTOMER");
DataObjectPtr root = readCustomers->executeQuery();
{panel}
Another piece of convention followed by the RDB DAS has to do with database
table keys. When asked to "apply changes", the RDB DAS scans the SDO
ChangeSummery and generates the set of INSERT/UPDATE and DELETE statements
needed to flush the graph changes to the database. To generate the correct
statements the RDB DAS needs to know how a DataObject's properties map to a
database table key. In typical Object/Relational mappings frameworks, this
information is provided in a config file with something like this:
{noformat}
<Table tableName="COMPANY">
<Column columnName="ID" primaryKey="true"/>
.
.
.
</Table>
{noformat}
However, the RDB DAS understands is that, in the absence of user provided
cofiguration data, a DataObject property named "ID" maps to a key column named
"ID" in the database. If a developer follows this convention (naming key
columns "ID") then this key-mapping information need not be specified in a
configuration file. Convention Over Configuration strings are not case
sensitives, so "ID", "Id", "iD" or"id" will be mapped as a primary key.
The RDB DAS also has convention in place for mapping DataObject relationships.
If a database table (yyy) has a column named xxx_ID then the DAS will assume,
in the absence of config data to the contrary, that there is a one-to-many
relationship from table xxx to table yyy and that the key column for table xxx
is named "ID".
There is more "convention over configuration" coming. We are currently looking
into support for optimistic concurrency control conventions and other areas.
---------------------------------------------------------------------
CONFLUENCE INFORMATION
This message is automatically generated by Confluence
Unsubscribe or edit your notifications preferences
http://cwiki.apache.org/confluence/users/viewnotifications.action
If you think it was sent incorrectly contact one of the administrators
http://cwiki.apache.org/confluence/administrators.action
If you want more information on Confluence, or have a bug to report see
http://www.atlassian.com/software/confluence
---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
