You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@openjpa.apache.org by "Krzysztof (JIRA)" <ji...@apache.org> on 2010/06/08 19:37:15 UTC

[jira] Updated: (OPENJPA-1687) @Strategy and @ElementStrategy handling resets/obstructs SQL type set in the custom handler map() call.

     [ https://issues.apache.org/jira/browse/OPENJPA-1687?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]

Krzysztof updated OPENJPA-1687:
-------------------------------

    Description: 

openJPA advertises itself as an expandable framework and provides @Strategy (and undocumented) @ElementStrategy annotations to create mappings.

I would like to create support for already existing JDBC4 standard mapping primitive arrays mapping in i.e. Postgres:

java double[]-> SQL float8[]

Unfortunately it is not possible to make it work with a @Strategy as enhancer blindly checks if the field is a collection/array and refuses to do anything with it.
In the case like this a field should not be delegated to an external table but simply embedded - same way LOBs are, with different toDataStoreValue/toObjectValue pair as provided in the strategy below..

If we use @ElementStrategy, i.e.:

<class>
 @PersistentCollection
    @ElementStrategy("PostgresJPAArrayHandler")
    @ElementColumns({
        @ElementColumn(name="TDOUBLE")
    })

 private double tdouble[];
</class>

then enhancer accepts it. Schema synchronisation creates wrong DDL suppressing given SQL type and mapping it to SQL keyword ARRAY. 


Does this rather simple requirement is not covered by openJPA? 
Would be grateful for any hints, comments...




public class PostgresJPAArrayHandler
    extends AbstractValueHandler {

 public PostgresJPAArrayHandler()
 {
 
 }
     /**
	 * Serialization ID
	 */
	private static final long serialVersionUID = 1L;
	private static final PostgresJPAArrayHandler _instance =
            new PostgresJPAArrayHandler();

    /**
     * Singleton instance.
     */
    public static PostgresJPAArrayHandler getInstance() {
        return _instance;
    }

    public Column[] map(ValueMapping vm, String name, ColumnIO io,
        boolean adapt) {
        Column col = new Column();
        col.setName(name);        
        col.setType(JavaSQLTypes.SQL_ARRAY);
        col.setTypeName("float8[]");    //////<<<--------------------- this gets reset in mergeField or whereabouts
 
        return new Column[]{ col };
    }

    public Object toDataStoreValue(ValueMapping vm, Object val,
        JDBCStore store) {
    	//this has to be array of some integer or double for now
    	if (val instanceof double[]) {
    		
			try {
				Array darr = store.getConnection().createArrayOf("double", (Object[]) ( ArrayUtils.toObject((double[])val)));
				return darr;
			} catch (SQLException e) {
				// TODO Auto-generated catch block
				e.printStackTrace();
			}
			
		}

        return  null;

    }
    /** Convert from DB rep into OM representation */
    public Object toObjectValue(ValueMapping vm, Object val) {
        if (val == null)
            return null;
        return ArrayUtils.toPrimitive((Double[])val);
    }
    
}

  was:

openJPA advertises itself as an expandable framework and provides @Strategy (and undocumented) @ElementStrategy annotations to create mappings.

I would like to create support for already existing JDBC4 standard mapping primitive arrays mapping in i.e. Postgres:

java double[]-> SQL float8[]

Unfortunately it is not possible to make it work with a @Strategy as enhancer blindly checks if the field is a collection/array and refuses to do anything with it.
In the case like this a field should not be delegated to an external table but simply embedded - same way LOBs are, with different toDataStoreValue/toObjectValue pair as provided in the strategy below..

If we use @ElementStrategy, i.e.:

<class>
 @PersistentCollection
    @ElementStrategy("gaia.cu7.dal.PostgresJPAArrayHandler")
    @ElementColumns({
        @ElementColumn(name="TDOUBLE")
    })

 private double tdouble[];
</class>

then enhancer accepts it. Schema synchronisation creates wrong DDL suppressing given SQL type and mapping it to SQL keyword ARRAY. 


Does this rather simple requirement is not covered by openJPA? 
Would be grateful for any hints, comments...




public class PostgresJPAArrayHandler
    extends AbstractValueHandler {

 public PostgresJPAArrayHandler()
 {
 
 }
     /**
	 * Serialization ID
	 */
	private static final long serialVersionUID = 1L;
	private static final PostgresJPAArrayHandler _instance =
            new PostgresJPAArrayHandler();

    /**
     * Singleton instance.
     */
    public static PostgresJPAArrayHandler getInstance() {
        return _instance;
    }

    public Column[] map(ValueMapping vm, String name, ColumnIO io,
        boolean adapt) {
        Column col = new Column();
        col.setName(name);        
        col.setType(JavaSQLTypes.SQL_ARRAY);
        col.setTypeName("float8[]");    //////<<<--------------------- this gets reset in mergeField or whereabouts
 
        return new Column[]{ col };
    }

    public Object toDataStoreValue(ValueMapping vm, Object val,
        JDBCStore store) {
    	//this has to be array of some integer or double for now
    	if (val instanceof double[]) {
    		
			try {
				Array darr = store.getConnection().createArrayOf("double", (Object[]) ( ArrayUtils.toObject((double[])val)));
				return darr;
			} catch (SQLException e) {
				// TODO Auto-generated catch block
				e.printStackTrace();
			}
			
		}

        return  null;

    }
    /** Convert from DB rep into OM representation */
    public Object toObjectValue(ValueMapping vm, Object val) {
        if (val == null)
            return null;
        return ArrayUtils.toPrimitive((Double[])val);
    }
    
}


> @Strategy and @ElementStrategy handling resets/obstructs SQL type set in the custom handler map() call.
> -------------------------------------------------------------------------------------------------------
>
>                 Key: OPENJPA-1687
>                 URL: https://issues.apache.org/jira/browse/OPENJPA-1687
>             Project: OpenJPA
>          Issue Type: Bug
>    Affects Versions: 2.0.0
>         Environment: OS X, Linux
>            Reporter: Krzysztof
>
> openJPA advertises itself as an expandable framework and provides @Strategy (and undocumented) @ElementStrategy annotations to create mappings.
> I would like to create support for already existing JDBC4 standard mapping primitive arrays mapping in i.e. Postgres:
> java double[]-> SQL float8[]
> Unfortunately it is not possible to make it work with a @Strategy as enhancer blindly checks if the field is a collection/array and refuses to do anything with it.
> In the case like this a field should not be delegated to an external table but simply embedded - same way LOBs are, with different toDataStoreValue/toObjectValue pair as provided in the strategy below..
> If we use @ElementStrategy, i.e.:
> <class>
>  @PersistentCollection
>     @ElementStrategy("PostgresJPAArrayHandler")
>     @ElementColumns({
>         @ElementColumn(name="TDOUBLE")
>     })
>  private double tdouble[];
> </class>
> then enhancer accepts it. Schema synchronisation creates wrong DDL suppressing given SQL type and mapping it to SQL keyword ARRAY. 
> Does this rather simple requirement is not covered by openJPA? 
> Would be grateful for any hints, comments...
> public class PostgresJPAArrayHandler
>     extends AbstractValueHandler {
>  public PostgresJPAArrayHandler()
>  {
>  
>  }
>      /**
> 	 * Serialization ID
> 	 */
> 	private static final long serialVersionUID = 1L;
> 	private static final PostgresJPAArrayHandler _instance =
>             new PostgresJPAArrayHandler();
>     /**
>      * Singleton instance.
>      */
>     public static PostgresJPAArrayHandler getInstance() {
>         return _instance;
>     }
>     public Column[] map(ValueMapping vm, String name, ColumnIO io,
>         boolean adapt) {
>         Column col = new Column();
>         col.setName(name);        
>         col.setType(JavaSQLTypes.SQL_ARRAY);
>         col.setTypeName("float8[]");    //////<<<--------------------- this gets reset in mergeField or whereabouts
>  
>         return new Column[]{ col };
>     }
>     public Object toDataStoreValue(ValueMapping vm, Object val,
>         JDBCStore store) {
>     	//this has to be array of some integer or double for now
>     	if (val instanceof double[]) {
>     		
> 			try {
> 				Array darr = store.getConnection().createArrayOf("double", (Object[]) ( ArrayUtils.toObject((double[])val)));
> 				return darr;
> 			} catch (SQLException e) {
> 				// TODO Auto-generated catch block
> 				e.printStackTrace();
> 			}
> 			
> 		}
>         return  null;
>     }
>     /** Convert from DB rep into OM representation */
>     public Object toObjectValue(ValueMapping vm, Object val) {
>         if (val == null)
>             return null;
>         return ArrayUtils.toPrimitive((Double[])val);
>     }
>     
> }

-- 
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.