mardi 15 avril 2014

New data package: what is expected from server?

I'm studying the ticket app.

It dizzies at first but after a while you can easily follow and learn from it.

But, IMHO, is an unrealistic scenario:

Business applications normally deals with hundreds and thousands of records not only few tens.


Nevertheless I have no clue on how server should send data to the application.

Lets look at the fields of ticket model:



Code:



fields: [
'status',
{ name: 'projectId', reference: 'Project' },
{ name: 'creatorId', reference: 'User' },
{ name: 'assigneeId', reference: 'User' },
{ name: 'created', type: 'date' },
{ name: 'modified', type: 'date' },
{ name: 'modifiedById', reference: 'User' }, {
name: 'statusName',
depends: 'status',
convert: function(val, rec) {
return this.owner.getStatusName(rec.get('status'));
}
}
],

Most of them are FKs.

Do I have to send from server only the columns specified by the model or do I have to include (nested) each referenced entity? (No idea how the JSON would be in the latter case...)

If former, do I have to load ALL the referenced entities to memory?

I mean, all Projects and Users?

What if we have 1000 Users?


On the other hand:

I just crafted an application which uses a products catalog which has +/-40,000 items. If I have some reference to such entity (Product) from other models (say OrderDetail) do I have to keep the 40,000 items in memory?


Seems that the "magic" of the new data packages just wont work on large datasets


Regards.





Aucun commentaire:

Enregistrer un commentaire