I suspect the byte[] result is a Java-serialized form of the target ID (serialized Integer or Long??). Not sure why it does it in this case. Could be a bug. But as you have guessed, this API behavior is already different in M6. Per CAY-2255 it will return a full entity [1]. So probably unsafe to rely on the current behavior (broken or not) anyways.
> I'd still like to stay on 4.0M5 for production use. Understood. But note that M6 (likely to be renamed to B1) is quickly approaching. Could be a matter of weeks. Andrus [1] https://issues.apache.org/jira/browse/CAY-2255 > On May 4, 2017, at 5:05 PM, Musall, Maik <m...@selbstdenker.ag> wrote: > > Hi, > > I have this query: > > ColumnSelect<Object[]> select = ObjectSelect > .query( PDCMarketingInfo.class ) > .where( someExpression ) > .columns( PDCMarketingInfo.CUSTOMER, Property.COUNT ) > .select( context ); > > where PDCMarketingInfo.CUSTOMER is a to-one relationship to another entity. > Each row returned is an Object[] of length 2, with the second being a Long as > expected, but the first is a byte[6]. What is this byte array? If I convert > it to Integer, I can't make sense of the resulting number as well. > > I know there is a newer API with .columnQuery() in the works, and although I > would be interested if that is a solution to this, I'd still like to stay on > 4.0M5 for production use. > > Thanks > Maik >
