A little feedback on the ColumnSelect API:
What little I’ve used it, I’m loving it so far. But there’s one thing that
bothers me quite a bit: If I specify a relationship property as a “column” when
fetching, the resulting values will be the primary keys of the related objects.
Wouldn’t it be more principle of least astonishment-y if these would instead be
actual objects/faults?
In other words, I’d basically expect this:
ObjectSelect
.query( Person.class )
.columns( Person.NAME, Person.DIVISION )
.select( someContext );
…to return the same result as this EJBQL:
new EJBQLQuery( "select p.name, p.division from Person p" );
Once I’ve used ColumnSelect a little more, I’ll probably have more feedback,
I’m a naggy guy :).
Cheers,
- hugi
> On 14. feb. 2017, at 14:17, Andrus Adamchik <[email protected]> wrote:
>
> In the two short months since M4 we already have a number of shiny new things
> in Cayenne:
>
> * New ObjectSelect API (selecting individual columns, functions, GROUP
> BY/HAVING), which is easily one of the major features of 4.0 as a whole.
> * JCache integration.
> * Auto-loading DI modules.
>
> Plus as many of the breaking changes as possible got pushed, such as Maven
> plugin renaming. Would be nice to get some feedback from early adopters on
> those until the Beta API freeze. So I suggest we start the M5 vote shortly.
> Nikita still wanted to clean up a few things in the new code. Once he's
> ready, hopefully in about a week, we can do the assembly and start the vote.
>
> And of course it would be great if we could maintain a similar pace going
> forward, which will depend both on development progress and PMC members
> availability to review the releases. Will see...
>
> Andrus
>
