RE: [CDBI] ChildOf
[prev]
[thread]
[next]
[Date index for 2005/09/10]
> This sounds useful, but I don't quite understand from your
> description how it
> differs from a has_a relationship from the child to the parent
> (other than it
> allows you to specify the accessor name for the parent). What am
> I missing?
>
> I implement similar parent/child relationships like so:
>
> Parent->columns('Primary' => qw/parent_id/);
> Parent->has_many(children => Child);
> Child->columns('Primary' => qw/child_id/);
> Child->columns('Others => qw/parent_id/);
> Child->has_a(parent_id => Parent);
>
> How does ChildOf simplify or improve on this?
>
> Thanks,
> Ed
>
Good point. Forgot to explain. The problem with has_a is that it takes
your primary column, your object id (that you are probably using everywhere
in urls, keys, etc) and turns it into a parent object. That's not very nice
from two perspectives:
#1. I like my id where it is! :-) I mean, really - when I say <a
href="/?a_id=<%$link->link_id%>"> I expect a number to show up. (I believe
it actually will - because of stringify - but read on).
#2. I don't want to go to the database for the parent just because I am
accessing the child's id.
OK. So #1 actually works as-is. Unless you have a custom stringify, your
$link->link_id in a string context should show up as a number. but it does
so by first going to the database , and then converting the result into the
number (the #2 issue). What's worse is - I may choose to override
stringify - and break a completely independent module - that's not so good.
All that said, has_a certainly has a place. It is what gave me the idea,
and I used it heavily to figure out how to make my code work.
I am looking forward to any disagreements to the arguments above. I am very
new to cdbi (shame on me) and may miss something important.
_______________________________________________
ClassDBI mailing list
ClassDBI@xxxxx.xxxxxxxxxxxxxxxx.xxx
http://lists.digitalcraftsmen.net/mailman/listinfo/classdbi