# The active record of ember data

**URL:** <https://discuss.emberjs.com/t/the-active-record-of-ember-data/2070>\
**Category:** Ember Data\
**Created:** [August 7, 2013, 5:14pm UTC](https://discuss.emberjs.com/t/the-active-record-of-ember-data/2070 "2013-08-07T17:14:56Z")\
**Posts on this page:** 1\
**Page:** 1

<div class="post-metadata">

**Author:** ![eccegordo](https://sea1.discourse-cdn.com/flex019/user_avatar/discuss.emberjs.com/eccegordo/32/14927_2.png) [@eccegordo](https://discuss.emberjs.com/u/eccegordo)\
**Post date:** [August 7, 2013, 5:14pm UTC](https://discuss.emberjs.com/t/the-active-record-of-ember-data/2070/1 "2013-08-07T17:14:57Z")

</div>

I am trying to get my head around querying into a data model both before and after it has been loaded from the server.

Part of my model is defined by a single table with several columns and reasonably large number of rows (above 200,000). Effectively read only.

I understand how to get stuff by ID. But what about the case where I have fetched most if not all this data before hand into memory and javascript? By what patterns can I dynamically query it? e.g. specific columns or sets of columns.

Any meta programming techniques or generic computed properties?

I suppose the other way is just to always fetch from from the server and let my backend (rails) figure out how to find the record. In this case am I better off just sending back an array of IDs or should I respond with the ids + the json objects of the records. Will ember data know how to place them in the local model graph? I assume this is where the identity map comes into play.

Do I search locally first and then go to the server when necessary?

What does a typical access pattern look like here?