Help with database structure

["Kimmo Myllyviita (LMF)" <Kimmo.Myllyviita@lmf.ericsson.se>]

> Here's an imagined case, which corresponds to my real situation:
> 
> Florist gets flower deliveries all the time. He wants to store 
> information about flower packets in the LDAP database.
> The packets are labeled something like this:
> 
> packet for Smith:
> packet=weddind:roses=red,daisies=10,pinks=20,price=100
> 
> The thing is that LDAP should add new packet only if it differs from
> already stored
> ones. (let's say that price doesn't matter in matching)
> so how does he organize the database?? would it have to be something like
> this:
> 
> dn: packet=wedding:roses-red:daisies-10:pinks-20,sn=Smith
> price:100
> 
> dn: packet=funeral:roses-red:daisies-50,sn=Smith
> price:120
> .
> .
> 
> [or what ever characters are allowed in order to separate the objects??]
> 
> now before adding, he can operate 
> ldapsearch -b "packet=wedding:roses-red:daisies-10:pinks-20,sn=Smith"... 
> to check if packet already exists.. right?
> 
> Question:
> Can anyone think of any other reasonable way to organize the base?
> Is it possible to make just "sn=Smith" as dn and put packet objects as
> attributes and somehow still relate these attributes to a specific
> packet???
> (Also remembering the matching criteria- being able to look for a match
> with 
> only one ldapsearch.) 
> I would appreciate any suggestions. Thanks!
>