You must define a Primary Key class of your own.
I'd suggest to take the 'PrimaryKey Example - Multiple Field' from the
Datanucleus Guide [1] as a starting point, but improve the marshalling logic
such that arbitrary strings within your PK columns are possible:
I believe, using ',' as the separator should not conflict with other logic;
private final static String SEPARATOR = ",";
/**
* Constructor accepting same input as generated by toString().
*/
public ComposedIdKey(String value)
{
StringTokenizer token = new StringTokenizer (value, SEPARATOR);
this.targetClassName = token.nextToken();
this.field1 = new String(Base64.getDecoder().decode(token.nextToken()));
this.field2 = new String(Base64.getDecoder().decode(token.nextToken()));
}
..
public String toString ()
{
// Give output expected by String constructor
final String left = Base64.getEncoder().encodeToString(this.field1);
final String right = Base64.getEncoder().encodeToString(this.field2);
return this.targetClassName + this.field1 + SEPARATOR + this.field2;
}
[1] http://www.datanucleus.org/products/accessplatform/jpa/mapping.html#identity
On 2018/06/21 20:09:20, Brian K <[email protected]> wrote:
> Hello,
>
>
>
> I’m modeling my domain objects over an existing sql server database. One
> of my tables has a primary key that has two columns of type char(10). The
> values that come from columns like this appear with trailing spaces. The
> OID generated by Isis for a single column primary key works fine and shows
> urls such as
> http://localhost:8080/wicket/entity/dbo.courtcd:s_3AN-P%20%20%20%20%20 .
>
>
> When I have a table that has two fields like this as the primary key, I get
> this error:
>
> Could not parse OID
> 'dbo.crtprecx:domainapp.modules.simple.dom.impl.CourtPrecinct_PK_3AN-P
> :ANCHORAGE '; should match pattern:
> ^((([!*])?([^:~$\^#]+):([^:~$\^#]+))((~[^:~$\^#]+:[^:~$\^#]+)*))([$][^:~$\^#]+)?([\^](\d+):([^:~$\^#]+)?:(\d+)?)?$
>
> Do I have to now create a primary key class to use the trimmed values
> of the string fields in the primary key, or is there another way to
> handle this?
>
> Thank you!
>
> Brian
>
