WEBVTT 00:00.140 --> 00:02.720 Let's talk about the generator functions. 00:02.750 --> 00:05.330 Start with a simple function. 00:05.330 --> 00:13.820 That's job is to generate a couple of random numbers so we can iterate over them and to return an array. 00:14.300 --> 00:15.440 So how can we do that. 00:15.440 --> 00:20.480 Well we need a result variable so we can keep all the elements in the array. 00:21.050 --> 00:25.130 Then we can use a loop one of the loops. 00:25.130 --> 00:31.010 So in this case we just pass an argument of how many elements do we want. 00:31.040 --> 00:33.380 And we just count down. 00:35.060 --> 00:40.340 Next up we need to generate the random number and add it to the array. 00:41.120 --> 00:45.710 And only then we can return this result this resulting array. 00:45.740 --> 00:53.480 So later on we can for example call this function in a for each loop and then iterate over the results. 00:54.230 --> 00:57.710 Now let's see how can we do that in a generator. 00:57.710 --> 01:06.080 Way to first simplify this countdown function, and also to save as much memory as possible. 01:06.590 --> 01:07.850 Right. 01:08.030 --> 01:10.490 So the first change is the return type. 01:10.490 --> 01:14.420 Generators return the generator type. 01:14.660 --> 01:17.630 Now we can proceed with simplifying this. 01:18.110 --> 01:21.260 So we get rid of this result variable. 01:21.260 --> 01:24.110 We don't need to store results anywhere. 01:24.110 --> 01:31.190 And the difference between the generator function and normal function is you don't use the return keyword, 01:31.190 --> 01:35.810 you use yield and everything else stays the same. 01:35.810 --> 01:44.360 So the way it works is now the countdown function would return something every single time you call 01:44.390 --> 01:45.080 yield. 01:45.080 --> 01:48.680 We don't have to wait for the whole array to be created. 01:48.680 --> 01:50.600 Thus we can save memory. 01:50.630 --> 01:54.650 We don't have to create a data structure to return it. 01:54.650 --> 02:00.020 First, we can yield something from the function as we go. 02:00.920 --> 02:03.200 Now let's do the practical part. 02:03.650 --> 02:09.260 Create a generator PHP and a starting tag. 02:09.260 --> 02:17.090 And let me start with some code already so we can have this countdown function return the traditional 02:17.090 --> 02:20.330 way and R for each loop. 02:20.330 --> 02:28.310 So to see the difference between this traditional function and a generator, I need to echo something 02:28.310 --> 02:30.590 when we generate a new number. 02:32.870 --> 02:37.610 So I'm going to say generating number and add a new line. 02:39.140 --> 02:46.730 And also when we echo something when we echo the number before we do that I'm also going to say something. 02:46.730 --> 02:50.060 This would be echoing number. 02:50.510 --> 02:52.190 Now let me run this. 02:54.290 --> 02:56.150 So you see how it works. 02:56.210 --> 03:01.130 The numbers are generated first the five numbers that we are generating. 03:01.130 --> 03:04.730 And only then they are echoed one by one. 03:04.730 --> 03:06.530 So obvious things. 03:06.530 --> 03:09.050 We need to keep those numbers someplace. 03:09.050 --> 03:10.820 So we need to use up memory. 03:10.850 --> 03:14.900 So this is a tiny example a very small array. 03:14.900 --> 03:21.620 But you can easily imagine processing hundreds of thousands or even millions of records. 03:21.620 --> 03:24.350 And every single time such code runs. 03:24.350 --> 03:31.210 And when you run PHP on servers, you might have multiple clients running at the same time, so you 03:31.210 --> 03:38.680 can see how we can inflate the memory usage if we need to process some huge amounts of data. 03:38.680 --> 03:42.280 And this is what generators are for. 03:42.850 --> 03:46.480 So let me first return a generator from here. 03:46.510 --> 03:52.870 Let's get rid of the result variable altogether. 03:56.140 --> 03:57.580 And instead. 04:00.640 --> 04:02.770 Let's yield the results. 04:02.770 --> 04:07.750 And now when I run it you immediately see the difference. 04:07.750 --> 04:11.590 So we generate the number and we instantly echo it. 04:11.620 --> 04:16.210 We generate another one, we echo, and so on and so on. 04:17.200 --> 04:26.800 So this means that you can say that this function returns multiple times until until it doesn't. 04:26.830 --> 04:33.790 So to summarize the generators are great if you are working with huge data sets and you'd like to be 04:33.790 --> 04:35.320 efficient with memory.