1
00:00:02,260 --> 00:00:11,260
Let's look into our friends in their raw form again and let's now say we want to transform exam scores

2
00:00:12,040 --> 00:00:17,620
to be an array where we only see scores higher than 60.

3
00:00:17,770 --> 00:00:22,840
Now this can also be done in the projection phase because I want to transform every single french record

4
00:00:22,840 --> 00:00:24,380
I don't want to group by anything,

5
00:00:24,460 --> 00:00:32,660
I just want to transform the array in there. And to achieve this, I'll add exam scores again, you could name

6
00:00:32,670 --> 00:00:34,190
this differently though,

7
00:00:34,190 --> 00:00:35,380
this is up to you

8
00:00:35,740 --> 00:00:40,570
and in there, you can now use the special filter operator.

9
00:00:40,600 --> 00:00:48,850
Now how does filter work? Filter allows you to filter out certain elements in an array and only return

10
00:00:48,850 --> 00:00:51,780
these that fulfill a certain condition,

11
00:00:52,000 --> 00:00:56,510
so here I would want to filter for the score being greater than 60.

12
00:00:56,530 --> 00:00:57,910
Now how does this work?

13
00:00:57,910 --> 00:01:05,550
First of all you need to add an input here and that is the array which you want to filter,

14
00:01:05,590 --> 00:01:10,480
in our case we could say it's obvious that I want to filter an exam scores because I will store the

15
00:01:10,480 --> 00:01:12,720
result in a field named exam scores

16
00:01:12,730 --> 00:01:16,210
but keep in mind, you can name this field however you want.

17
00:01:16,210 --> 00:01:22,320
So therefore filter needs to know where is the input array and the input array is exam scores with a dollar sign

18
00:01:22,330 --> 00:01:25,310
because I'm pointing at a field in my document.

19
00:01:25,570 --> 00:01:27,820
Then we assign a temporary name,

20
00:01:27,820 --> 00:01:32,590
this is a so-called local variable which will be used inside our filter expression which we'll write

21
00:01:32,590 --> 00:01:33,640
in a second,

22
00:01:33,640 --> 00:01:35,500
you can name this however you want,

23
00:01:35,500 --> 00:01:37,600
I'll name it sc to not repeat

24
00:01:37,630 --> 00:01:38,740
scores all the time

25
00:01:38,770 --> 00:01:42,940
but you can name this however you want. The last element we need

26
00:01:42,950 --> 00:01:48,460
or the last field is the condition, cond.

27
00:01:48,500 --> 00:01:54,590
Now the conditions here can take a bunch of expression operators to which you also find a link in

28
00:01:54,590 --> 00:01:59,630
the last lecture of this module and I'll just bring this into a new line to make it easier to read. The

29
00:01:59,630 --> 00:02:05,570
expression operator I want to use here is greater than, now greater than here works a bit different than

30
00:02:05,570 --> 00:02:11,390
when we used it in match or find. Here greater than simply takes an array of values it should compare

31
00:02:12,140 --> 00:02:20,150
and this makes sense because we are now in the context of another operator, so we can't compare exam

32
00:02:20,150 --> 00:02:22,950
scores the whole array or anything like that,

33
00:02:23,240 --> 00:02:31,790
instead here what I want to compare inside of my filter expression is of course the value of this temporary

34
00:02:31,790 --> 00:02:36,020
variable which will refer to the different values in my exam scores,

35
00:02:36,020 --> 00:02:42,550
so you can imagine this being a function that gets executed for every exam score in my exam scores array

36
00:02:42,830 --> 00:02:47,780
and then every value will temporarily be stored in here so that we can compare it.

37
00:02:47,840 --> 00:02:56,420
So here in greater than, I will check if my sc variable is greater than 60,

38
00:02:56,450 --> 00:02:58,110
this is how this works.

39
00:02:58,110 --> 00:03:01,790
However sc like this would be treated as a string,

40
00:03:01,850 --> 00:03:05,980
we need to tell mongodb that we're referring to this variable.

41
00:03:05,990 --> 00:03:09,940
Now one dollar sign would look for a field named sc,

42
00:03:10,310 --> 00:03:13,040
we add two dollar signs in this case,

43
00:03:13,040 --> 00:03:19,280
you don't see it that often because you don't have that many operators of this kind but filter which

44
00:03:19,280 --> 00:03:25,040
essentially executes a function over and over again on all the fields in exam scores or all the elements

45
00:03:25,040 --> 00:03:31,670
in there I should say, filter need such a temporary variable for every step it executes for us and we

46
00:03:31,670 --> 00:03:35,230
can refer to this temporary variable with the double dollar sign

47
00:03:35,280 --> 00:03:42,470
syntax then. Now with this approach, we would not get the result we want though because you have to

48
00:03:42,470 --> 00:03:49,690
keep in mind that exam scores is in the end and I can show this to you again, is in the end

49
00:03:49,700 --> 00:03:57,890
an array of embedded documents where you have a difficulty and score. Now sc here would refer

50
00:03:57,890 --> 00:03:59,920
to the overall embedded document,

51
00:04:00,170 --> 00:04:02,150
now I'm interested in the score though,

52
00:04:02,450 --> 00:04:10,040
so I have to check sc.score here in my greater than condition. And with that I can take that aggregate

53
00:04:10,310 --> 00:04:12,840
function, execute it

54
00:04:13,040 --> 00:04:18,940
and we should get arrays with only scores greater than 60,

55
00:04:19,010 --> 00:04:20,340
so you can see that here.

56
00:04:20,910 --> 00:04:23,840
I get one document with scores greater than 60,

57
00:04:23,840 --> 00:04:26,610
then we got another document with scores greater than 60,

58
00:04:26,620 --> 00:04:29,870
there we only got one exam, hence we only got one object in there

59
00:04:30,020 --> 00:04:32,130
and then this last one.

60
00:04:32,150 --> 00:04:34,030
So this is how filter works,

61
00:04:34,100 --> 00:04:41,270
filter allows you to filter arrays in documents inside of the projection phase.