[dojo-contributors] dijit.Tree assumes datastore implements Identity API

Brian Douglas Skinner skinner at dojotoolkit.org
Tue Aug 21 14:33:44 EDT 2007


Michael Smith wrote:
> While I wasn't involved in the initial decision to have a separate 
> Identity API, I was under the impression that this was done because 
> some datastores may have data that don't have IDs. 

I think the original goal was just to keep the basic Read API as simple 
as possible, to keep it easy to write new datastores.  But 
realistically, implementing the full Read API is now somewhat 
complicated, and adding in Identity support would be fairly simple for 
most stores.

> But is there any case where a datastore may need to read data 
> to which it can't apply Identities?

I think a datastore can probably always just make up a unique id for 
each item, even if there's no unique id embedded in the data itself. 
For example, a .csv file won't necessarily have an id column in the file 
itself, but our dojox.data.CsvStore still provides a getIdentity() 
implementation by just returning the row number of the item in the file.

:o) Brian






More information about the dojo-contributors mailing list