@codeverse/ember-data-has-many-query
v0.2.1
Published
Ember data addon for querying async has-many relationships using query parameters
Downloads
1
Readme
ember-data-has-many-query
Ember Data's DS.Store
supports querying top-level records using the
query
function. This provides support
for things like pagination and searching.
However, DS.hasMany
and DS.belongsTo
cannot be queried in the same way. This means pagination and searching are not
supported with has-many/belongs-to relationships.
This addon provides a way to query has-many and belongs-to relationships. Currently the DS.RESTAdapter
and the
DS.JSONAPIAdapter
are supported.
Installation
ember install ember-data-has-many-query
Usage
Add the RESTAdapterMixin
to your DS.RESTAdapter
(or DS.JSONAPIAdapter
) extension:
import HasManyQuery from 'ember-data-has-many-query';
export default DS.RESTAdapter.extend(HasManyQuery.RESTAdapterMixin, {
});
Add the ModelMixin
to any DS.Model
extensions:
import HasManyQuery from 'ember-data-has-many-query';
export default DS.Model.extend(HasManyQuery.ModelMixin, {
});
Models with the mixin now support has-many/belongs-to queries:
post.query('comments', { page: 1 });
Sticky belongs-to
Ember Data 2.3.x and below: each has-many query calls reload
on the relationship's DS.ManyArray
. This means that all previously
queried records are cleared from the array. If you are caching the records from each query separately
(for example, in a separate array for an infinite scroll implementation), the inverse belongs-to
relationship is also cleared on those cached records.
If you want to keep the associated belongs-to record after a new query, you can define the belongs-to
attribute using belongsToSticky
:
import HasManyQuery from 'ember-data-has-many-query';
export default DS.Model.extend(HasManyQuery.ModelMixin, {
post: HasManyQuery.belongsToSticky('post'),
});
This is a (pretty terrible) hack that caches the belongs-to record in a separate property, and when the
record is cleared by another query call, any property get
s will return the cached version instead. If
anyone has ideas for better implementations, please let me know!