I am working on my first angular directive and still getting my head around the concepts and what's possible with directives. As I've been researching the best way to tackle this problem I haven't been able to identify an example that addresses what I'm trying to do, so thought I would ask for some help from the experts here.
I have an array of objects that are one of three types.
I would like to use the ng-repeat directive to iterate through this array and display the objects on the page.
Each object type has a different view associated with it as each object shares some properties, but also have unique properties.
I would like to set up a directive that displays the correct view based on the objective type.
So the logic would work something like the following:
<div ng-repeat="item in dataset">
<the-smart-directive>item</the-smart-directive>
</div>
One idea would be to have one directive where I determine the templateUrl based on the object type and then have a unique template for each of the objects.
Another idea would be to have a parent directive and then three other directives (one for each object type) and the parent directive would insert the correct object type directive (this is the idea that seems like the better approach, but I'm not sure how to actually implement this idea).
I'd love some help in understanding the best way to tackle this and how to implement. If you could provide some example code that would be wonderful and get me started on the right path.
Thanks for your help!
The way we are using it is with ng-switch inside the ng-repeat.
<div ng-repeat="item in dataset" ng-switch="item.type">
<directive-one ng-switch-when="1">
</directive-one>
<directive-two ng-switch-when="2">
</directive-two>
<directive-three ng-switch-when="3">
</directive-three>
</div>
Related
I'm trying to go with the best approach and avoid unnecessary rendering/processing time in my AngularJS app when choosing between 2 directives to be displayed in the page inside an ngRepeat loop, want to know which is the best way:
If by setting the ng-if directly in the directive html element, like:
<div ng-repeat="element in list">
<my-directive-a ng-if="someFunction(element)"></my-directive-a>
<my-directive-b ng-if="!someFunction(element)"></my-directive-b>
</div>
Or by moving out the first <div> from the directive's template and use it as a wrapper for each directive. For instance:
<div ng-repeat="element in list">
<div ng-if="someFunction(element)">
<my-directive-a></my-directive-a>
</div>
<div ng-if="!someFunction(element)">
<my-directive-b></my-directive-b>
</div>
</div>
NOTE: The starting <div> element on each directive could be modified behave the same so I will basically take that out of the directive's html and moving it outside the directive declaration in order to place the ng-if there
What would be the best approach for this case? Are there any performance implications from doing it one way or another? Or is it just the same thing? Consider that the number of elements in the list could get really big.
They are quite the same, but you can improve performance with one-time binding, but only when element does not change at runtime (for example, let's say that it has property name, and your someFunction is like return element.name === 'John'). Angular just stop observing this function when it returns value, and watches will be deleted. There are 2 prerequisites to use this solution:
Elements properties in list does not change (if you rely on them in someFunction), for example if you rely on name property name must not change, because watcher on someFunction is note available.
When list changes or its elements properties change, you reload all list (for example, you fetch it from server again if you know that change occurred)
What you get with this? There is no watches after my-directives are drawn on ng-ifs, and when something changes, new reference is bound to list (for example, it comes from server) and everything will be redrawn, ng-ifs will run again and when will become stable (function returns value) then will be unbound. How it looks like? Like this:
<div ng-repeat="element in list">
<div ng-if="::(someFunction(element))">
<my-directive-a></my-directive-a>
</div>
<div ng-if="::(!someFunction(element))">
<my-directive-b></my-directive-b>
</div>
</div>
Two colons before expression. But be aware, that with one-time binding it's easy to mess up - you need to be sure that you test your code enough to be sure it works.
What is the advantage of using "Controller as" syntax in Angular? Is it just to create an alias for a controller or does it have some other technical reasons behind the scenes?
I am new to Angular and want to know more about this syntax.
controllerAs-syntax has multiple advantages:
Clartiy
Consider the following example:
<div ng-controller="containerController">
<h2>Improve your life!</h2>
<p ng-controller="paragraphController">
We talk about {{topic}} a lot, but do we really understand it?
Read this article to enhance your knowledge about {{topic}}
</p>
</div>
Just by reading this piece of code, you can not tell where topic comes from. Does it belong to the containerController, to the paragraphController or is it just a random floating scope variable from sone input above?
By using controllerAs it is very clear:
<div ng-controller="containerController as container">
<h2>Improve your life!</h2>
<p ng-controller="paragraphController as paragraph">
We talk about {{paragraph.topic}} a lot, but do we really understand it?
Read this article to enhance your knowledge about {{paragraph.topic}}
</p>
</div>
You can immediately see that topic is a property of paragraphController. This makes code a lot more readable overall, as it forces the developer to make clear who the functions and variables in the scope belong to.
Binding to properties
When you are using the old controller syntax, strange things can happen when you have multiple bindings in different scopes to the "same" variable. Consider this example:
<form ng-controller="myFormController">
<input type="text" ng-model="somefield">
<div ng-controller="someOtherController">
<input type="text" ng-model="somefield">
</div>
</form>
It looks like both inputs are bound to the same variable. They are not. Everything looks like it works fine when you edit the first input first, but as soon as you edit the second one, they won't sync up anymore. This has to do with the way scope-inheritance and binding work(and there is an excellent answer on this on SO). This does not happen when you bind to object properties (aka when there is a . in your ng-model-attribute). With controllerAs you bind to properties of the controller objects anyways, so it naturally solves that problem:
<form ng-controller="myFormController as myForm">
<input type="text" ng-model="myForm.somefield">
<div ng-controller="someOtherController as other">
<input type="text" ng-model="myForm.somefield">
</div>
</form>
It gets rid of scope(mostly)
The use of scope to create bindings to controllers in old angular code is hard to read, hard to understand and completely unnecessary if you use controllerAs. You no longer have to inject scope into every single controller, in fact you will probably not inject it in any in most applications (you still need to do so if you want to use the angular event system). This results in cleaner controller-code with less strange boilerplate.
It prepares for Angular 2
In angular 2, scope will be gone and we will write everything as components. Using controllerAs lets you work without scope and forces you to think more component-oriented, thus preparing you (and the applications you eventually will want to migrate) for the 2.0 update.
Several times when creating or customizing a directive (either my own directive or for example https://github.com/dpiccone/ng-pageslide) I get a point where all the display logic is controlled by a single css class. At that point the directive boils down to adding and removing a single class. So instead of using a new directive I can simply use the ng-class directive (see an example here: https://gist.github.com/Hypercubed/8f40556eb0f6eddbcca3). Is there an advantage to the custom directive approach vs the ng-class/CSS styles approach? I guess the custom directive doesn't depend on $animate? Am I just doing it wrong?
Sorry for another directive vs. XXX question.
I think you're failing to see the forest for the all tress. You're focusing on a very minute detail and missing the larger picture. Directives are more than simply applying styles. I think an example is best. For example, take the rating directive. If you wanted to render a star rating model it might look like this:
<div ng-rating="album.rating" max="5"></div>
That may add the following to the DOM:
<ul class="inline">
<li ng-repeat="i in max">
<i ng-class="{ 'icon-start-empty': i > rating, 'icon-star': i <= rating }"></i>
</li>
</ul>
Under the covers ng-class it utilized, but that is only a part of the logic encapsulated in the rating directive. It allows the user to configure how many stars the rating is out of, and renders the same number of li elements. Then because you wrote a directive it allows you to reuse this logic where ever it's required. Using ng-class only works in that 1 location. If you want to do the same thing you're copying code which is a sign maybe you want to wrap that logic up in a directive.
I have no idea how to pre-populate an ng-model in this circumstance that I have to use ngBind. I tried ng-init, but it's not working.
<h6 ng-show="isOwner" ng-bind="currentMagazine.magazine_name"
contenteditable ng-model="currentMagazine.magazine_name"
ng-change="update()"></h6>
I have a seperate directive that binds contenteditable attributes to ngModelController.
The problem now is whenever I update the model, ng-bind will jump out and refresh the div element, resulting in the cursur going back to the beginning of the text, which makes it impossible for any user to type.
I tried ng-init in a fashion like this:
<div ng-init="magazineName = currentMagazine.magazine_name">
<h6 ng-show="isOwner"
contenteditable ng-model="magazineName"
ng-change="update()"></h6>
</div>
It's not working. If I don't use ng-bind, then no text will show up.
Also I notice it might be related to this problem, when I type with space or delete key, they are escaped into HTML entities...so you get result like this:
Hopefully both of my problems can be solved! Thank you (it's a very frustrating day)!
In your app.js do this:
$scope.magazineName = $scope.currentMagazine.magazine_name;
In your HTML do something like this:
<input
ng-show="isOwner"
contenteditable
ng-change="update()"/>
or
<h6 ng-show="isOwner"
contenteditable
ng-model="currentMagazine.magazine_name"
ng-change="update()">
{{currentMagazine.magazine_name}}
</h6>
... though perhaps not, it's a bit difficult to gauge without seeing it... please make a jsfiddle or plunkr if you'd like to get more eyes on it.
If you're just trying to make some large text that is still editable and bound to the model it may be easier to just style an input for your needs.
Decided to play with contenteditbale since it's chance for me to learn something too... I can't seem to recreate the issue though:
http://jsfiddle.net/VSJQX/
I saw after doing this it wasn't updating the model, found another SO post that resolves that and included the changes here:
http://jsfiddle.net/VSJQX/2/
Is there a way to have a HTML-view with pre-populated values from the server, and then get AngularJS to read those values into it's $scope?
I'm thinking of a scenario where the HTML is like this:
<div ng-controller="TestController">
<div ng-bind="title">Test Title</div>
<div ng-bind="itemCount">33</div>
<div ng-repeat="item in items">
<div ng-bind="item.title">Item 1 Title</div>
</div>
</div>
<button ng-click="update()">Update</button>
And the JavaScript is like this:
function TestController($scope) {
$scope.update = function() {
console.log($scope.title); // Should log "Test Title"
};
}
The thought behind this is to let the server render HTML that search engines can index, but have a JavaScript-model-representation of the content for manipulation through JS.
While ng-init is one solution, it requires you to explicitly set the value. So here is an alternative solution.
http://plnkr.co/edit/pq8yR9zVOHFI6IRU3Pvn?p=preview
Note : This solution wont work for ng-repeat. Control flow directives cant be used with this. But for simple extraction of information from ng-bind this works pretty well. All that you need to do is add the default directive ( code in plunk ) to wherever you are doing the bind and it will extract the text content and push it to the scope variable.
EDIT (solution with ng-repeat):
So, I was thinking of a way to make ng-repeat also work the same way. But getting ng-repeat to work like this isnt an easy job ( see the code for proof :P ). I have finally found a solution - here you go :
http://plnkr.co/edit/GEWhCNVMeNVaq9JA2Xm2?p=preview
There are a couple of things you need to know before you use this. This hasnt been thoroughly tested. It only works for repeating over arrays ( not objects ). There could be cases that have not been covered. I am overriding ngRepeat itself which could have other consequences. When you loop through the items ( in your server side code ) dont forget to add default="true" on the first element and default on the rest of the elements.
Hope this helps.
Add ng-init to your elements with the value so that it will work the way you want.
http://docs.angularjs.org/api/ng.directive:ngInit
I think what you really want is to make your application searchable by serving static files in parallell. Read more about it here http://www.yearofmoo.com/2012/11/angularjs-and-seo.html