Install and Configure D3 v4

Convert Input Data to Output Values with Linear Scales in D3 v4

Convert Dates to Numeric Values with Time Scales in D3 v4

Create Labels from Numeric Data with Quantize Scales in D3 v4

Create Labels from Non-numeric Data with Ordinal Scales in D3 v4

Load and Inspect Data with D3 v4

Select DOM Elements with D3 v4

Modify DOM Elements with D3 v4

Create DOM Elements with D3 v4

Start Visualizing Data Driven Documents with D3 v4

 Output SVG Elements with D3 v4

SVG Graphics Containers and Text Elements in D3 v4

Basic Interactivity with D3 v4

Better Code Organization with selection.call() with D3 v4

Margin Convention with D3 v4

Create Chart Axes with D3 v4

Make D3 v4 Charts Responsive with the viewBox attribute

Build a Column Chart with D3 v4

Build a Scatter Plot with D3 v4

Build a Line Chart with D3 v4

Build an Area Chart with D3 v4

Debug D3 v4 with Dev Tools

Animate Transitions in D3 v4

Reuse Transitions in D3 v4

Animate with the General Update Pattern in D3 v4

Animate Chart Axis Transitions in D3 v4

Build Interactive JavaScript Charts with D3 v4

In D3, the General Update Pattern is the name given to what happens when a data join is followed by operations on the enter, update, and exit selections. When a chart's data changes over time and each update can both create new elements and destroy existing ones, the General Update pattern can help convey meaning to users. This lesson demonstrates the pattern using animated transitions on a column chart.

https://app.egghead.io/api/v1/lessons/d3-animate-with-the-general-update-pattern-in-d3-v4/subtitles

[00:00] In this example, we have a basic column chart where we're plotting the test scores of five different people. We just have a simple HTML page here. We've got our chart div that things get drawn into. We're specifying the fill style of our rectangles here.

[00:16] If we go take a look at our code, we can see that our data is just an array of objects that each have a name property and a score property. We're then using the D3 margin convention and specifying the margin's width and height.

[00:30] Then from that, creating our SVG tag and the container graphics element. We're calling responsivefy so that everything stays responsive when our window resizes. Then we go on to creating our scales. Our X scale is a band scale.

[00:49] Since we're creating a column chart here, that's what we want to use. Our domain for our X scale is just the list of name properties from our data. We're just pulling those names out, and using that to create our X scale so that we have everybody's names below their column.

[01:06] We've got a little bit of padding so that we've got some space between our columns there. Then we attach that bottom axis to our chart. Our Y scale is just hard-coded to run from 0to 100 for its domain. Since they're test scores, we're just going to say it's 0to 100. We don't need to examine the data at all.

[01:29] Finally, we go ahead and create the rectangles by doing select all rect, doing our data join, where we pass in the array of data objects. Then once we get to our enter selection, we append a rectangle. We set the X and Y based on our X and Y scales.

[01:49] For our width, we're able to use the bandwidth method that is unique to band scales. That takes care of sizing our columns horizontally. Then for our height, we're again just using the score passed to our Y scale, and subtracting that amount from our height.

[02:09] This is a nice functional chart, but it only renders once. Nothing changes over time. What we're going to do is look at the general update pattern, and how we can accommodate data that does change over time.

[02:23] The first thing we're going to do is we're going to come over here to our HTML and add back some buttons that I created that each one of them just has an on-click handler that is going to call a method named render, and pass in a subject name.

[02:37] I'm going to save that. We need to make a little bit of room for that in our chart. We're going to shrink this a little bit. Everything's laid out there. Now I'm going to paste in some data that is same basic structure, but instead of just a score property, we actually have math, science, and language properties.

[02:58] You'll also notice that some of the people have a null value for some of the subjects. If a person has a null value, they won't be shown on the chart. All of this stuff with our margins, and our axes, and our scales, all of that can stay the same.

[03:16] What we're really concerned with updating is this block of code here. This is what we're actually going to turn into our render function. We're going to say render. It's going to take a subject, and then we'll paste our code in here.

[03:32] If I give this subject parameter a default value of math, and then I'm going to go update these. Instead of reading a score property, we're going to use that subject parameter and read that property name off of the data object.

[03:49] Do the same thing down here, change that to subject. Then we'll go ahead and add a default call to render, which will use that default value of math. Now if we save this, we get the same data, except Billy is no longer there, because in our new data, Billy has a null value for math.

[04:12] The first thing we actually want to do is handle those null values. It automatically handled that for us here because that value is missing. What we actually want to do is, when we call this method with a subject, we want to filter that array so that we only get back items that have a value for that subject.

[04:35] For this, we're just going to use a simple array filter method, and say data.filter. Then when we get each item, we'll check for the presence of that subject. This is essentially saying if the math property, in the default case, has a truthy value, then it will be included. If not, it will not be included.

[04:57] In order to handle things separately, we need to break out our different selections. If you'll remember, our default value that gets returned from a data join is the update selection. Essentially, these two lines here, it selects all of the rectangles that may be there, joins the data, and that gives us our update selection.

[05:21] We get our enter selection from our update selection. Update.enter is where we're going to append the rectangles. Before we get to that enter selection, let's deal with the exit selection. Remember, the exit selection is where you have a DOM element whose corresponding data item is not in the data.

[05:46] If we were to say update.exit.remove, that is going to remove any of the rectangles who don't have a corresponding data object in whatever data has been joined to the update selection. That will get rid of elements that we no longer need.

[06:09] Any elements that we don't have anything for, so we're starting off here. Billy has a null value for math, and so he doesn't have an element. When we switch to science, where we does have a value, we're going to need to create a rectangle for him.

[06:25] We can actually go ahead and cut this off here. We'll assign this to a variable named enter. Our enter variable is our enter selection. We're going to append a rectangle for everything in that. Then all of this code that actually sizes and positions everything, we want that to apply to our update and our enter selections.

[06:47] To do that we're going to say update.merge, and then pass in our enter selection. This is creating a selection that is both our update and our enter selection merged together. If we save this, we get math re-rendered here.

[07:05] If we click science, we see Billy does appear, and David gets removed. That is because David does not have a value for science. If we look at language, we see that Cindy is the one that doesn't have a value.

[07:21] If we click language, Cindy disappears, and David comes back. Now we have a chart that will actually update based on whichever subject we tell it to render. This is nice, and it's entirely functional, but it would be even better if we could convey what was going on a little bit through animations.

[07:43] That's what we're going to do now. If we come down to our render function, the first thing I'm going to do is I'm going to create a transition that we can use throughout here. We're going to say d3.transition.duration, and I'm going to give it a one second duration.

[08:02] Sometimes it's helpful to use a little bit slower of a duration when you're just working through things, so you can tell what's going on. The first thing that we want to do when the subject changes is we want to animate the items that are being removed out of the chart.

[08:20] We've got our exit selection here, but instead of removing things immediately like that, we're actually going to say transition, and pass in the T instance there. We've got the duration already defined for us. We're going to set the Y attribute of those rectangles to be the height of our chart.

[08:42] We're going to set the height attribute of those rectangles to be zero. They're going to simultaneously move down and shrink so that we get those nice slide out animations. When that slide out animation is done, then the element will finally be removed.

[09:03] After the elements that are being removed have been transitioned out, then we want to update any of the items that are remaining on the chart. To do that, we'll just work with our update selection. We're going to say update, and again, .transition, and pass in our transition instance.

[09:23] We're going to add a delay this time of one second, because we want this to happen after our exit transition. Actually, we'll just copy this code from down here, since this is what we want to happen. We're going to specify that the Y gets updated and the height gets updated.

[09:49] There we've got that defined. Lastly, we just need to deal with our enter selection. We're going to slide everything out. We're then going to update things that are still there. Then we're going to slide in any new items.

[10:03] To do that, we're going to go ahead, and we don't actually need this enter variable anymore. We don't even need this selection merging anymore. We're going to say take the enter selection, append the rectangle, but when it's initially appended, we actually want to use these same properties that we're using that we're transitioning out of.

[10:30] We want things to start there for our new items. We'll paste that down here. The new rectangles are going to start at the bottom with no height. We can get rid of this line here, and this line here. Our X and our width can be set to their final values initially. There's no problem with that.

[10:51] Then we can actually go ahead and do our transition again here. This time, we're going to set our delay to be two seconds, because remember, we've got one second for our exit transition, one for our update. That's two full seconds that we need before we transition things in.

[11:11] Now we can copy these lines down here, because that's where we want things to end up for this as well. Now if save this, things update, and then they slide in. If we click on science, it's not actually working. Why is that?

[11:31] Aha, in order for all of this to work, each of these elements has to be associated with the same data object, regardless of what subject is being rendered. To do that, we need to use a key function when we call the data function here.

[11:49] What we need to do is define this method. It's again just going to be a little arrow function that gets each item passed in. We're just going to return the name. Essentially, what this is saying is when you do this data join, we want you to look at this name property to uniquely identify each one of these objects.

[12:13] If we look at our data, the name property is unique, and we're just changing which subject we're looking at. We need that key function so that it knows which rectangles go with which data objects. If we save this, we've got to refresh. We've got a little bit of a delay there.

[12:31] If we switch to science, David slides out, everybody updates, and Billy slides in. Go to language, Cindy slides out, update, David slides in. This is great. This is working perfectly. If we refresh, we see a little bit of a problem, though, because we've got that delay at the beginning.

[12:50] That's because we're delaying our enter animation by two seconds, and that doesn't make sense to do on the first render. What we can actually do is make this conditional. We can say update.exit.size. If it's true, we'll use a delay of two seconds, if not, it's zero.

[13:14] This is essentially saying if you have any items to remove from the chart, go ahead and implement that two second delay. If you don't, don't delay things at all. This isn't production quality code, because you may have an empty exit selection, but already have some items on there in your update selection.

[13:35] For our purposes, this gets the point across. Now if we save this, it is immediately going to slide those in initially, then we can switch subjects, and we see a column slide out. Everything else updates, and then the new column slides in.

Animate with the General Update Pattern in D3 v4

Share this video with your friends

Course

Build Interactive JavaScript Charts with D3 v4