I'm fully in favor of schema-qualified table names, but I'd much rather see a more dynamic approach. We've got a number of identical schemas that we use. It would be much more useful to us here if the schema were able to be assigned at runtime rather than at compile time.
-----Original Message----- From: Michael Kashambuzi [mailto:[EMAIL PROTECTED] Sent: Friday, May 28, 2004 2:34 PM To: [EMAIL PROTECTED] Subject: Changes to Support Fully Qualified Tables Names Torque Users and Developers: I find Toruqe to be one of the best and easiest to use persistence layers but was very dimayed by its lack of support for fully-qualified table names. So I decided to make a few changes, and in the spirit of open source, share those changes. I ran all the existing Torque tests so I feel comfortable in making the JARs available to anyone who is interested. I made the following changes to the torque-gen-3.1 source code to support fully qualified table names: 1. TorqueJDBCTransformTask.java. now generates schema.xml with attributes name="@[EMAIL PROTECTED]" javaName="TableName 2. build-torque.xml now includes an ant task for filtering @SCHEMA@ from schema.xml when generating ${project}-schema.xml with ${torque.database.schema} specified in build.properties and the following in torque-3.1 source code to support fully qualified table names: 1. BasePeer.java now determines the table names using columnName.lastIndexOf('.') orderByColumn.lastIndexOf('.') join#.lastIndexOf('.') instead of using columnName.indexOf('.') orderByColumn.indexOf('.') join#.indexOf('.') 2. Criteria.java now determines table name using int dot = tableColumn.lastIndexOf('.'); instead of int dot = tableColumn.indexOf('.'); Here is a sample of the relevant snippets from schema.xml <?xml version="1.0"?> <!DOCTYPE database SYSTEM "http://db.apache.org/torque/dtd/database_3_1.dtd";> <!-- Autogenerated by JDBCToXMLSchema! --> <database> <table javaName="MyTable" name="@[EMAIL PROTECTED]"> ... </table> </database> and the relevant snippets from ${project}-schema.xml <?xml version="1.0"?> <!DOCTYPE database SYSTEM "http://db.apache.org/torque/dtd/database_3_1.dtd";> <!-- Autogenerated by JDBCToXMLSchema! --> <database> <table javaName="MyTable" name="MYSCHEMA.MY_TABLE"> ... </table> </database> and the relevant snippets from BaseMyTablePeer.java public abstract class BaseMyTablePeer extends BasePeer { /** the default database name for this class */ public static final String DATABASE_NAME = "default"; /** the table name for this class */ public static final String TABLE_NAME = "MYSCHEMA.MY_TABLE"; /** the column name for the MY_COLUMN field */ public static final String MY_COLUMN; static { MY_COLUMN = "MYSCHEMA.MY_TABLE.MY_COLUMN"; } ... } I also made an additional change which allows the user to specify a property, ${torque.database.tables} = ${databaseTables}, which allows Torque to generate files only for a subset of tables. 1. TorqueJDBCTransformTask.java has an additional attribute, dbTables, which is set from ${torque.database.tables} and only includes this subset of tables when generating the schema.xml file. To include all tables, leave this blank or you can set it to %. Lastly, I found bugs which did not correctly define or provide casting for default values for primitive values. Torque was doing private short someNumber = ; instead of private short someNumber; and, in the method protected MyTable copyInto(MyTable copyObj) throws TorqueException, Torque was doing copyObj.setSomeNumber(0); instead of copyObj.setSomeNumber((short)0); so I updated in Object.vm to account for these. Again, if you'd like a diff of the source or the generated JAR files, I can gladly provide these. Michael --------------------------------------------------------------------- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED] --------------------------------------------------------------------- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
