1
00:00:02,230 --> 00:00:07,620
So you learned a lot about projection and that there are many many many operators you can use for that

2
00:00:07,880 --> 00:00:12,690
in the last lectures, which is super important because that is one of the core things

3
00:00:12,880 --> 00:00:16,120
the aggregation pipeline is all about.

4
00:00:16,150 --> 00:00:21,760
Now let's project a bit more because this really is an important step of almost any pipeline you're

5
00:00:21,760 --> 00:00:22,770
going to build

6
00:00:23,110 --> 00:00:26,020
and for that, let's first of all have a look at our input data

7
00:00:26,020 --> 00:00:26,880
one more time.

8
00:00:26,950 --> 00:00:31,470
Thus far, we basically only included the gender and our transformed name here,

9
00:00:31,810 --> 00:00:33,500
now what else could we include?

10
00:00:34,390 --> 00:00:41,700
Well for example, the location, we could already prepare the location so that we can later work with it,

11
00:00:41,710 --> 00:00:49,010
so what we could do is we could basically turn the location into a geo json object where we have a

12
00:00:49,010 --> 00:00:54,170
type which will be a point because we have two coordinates and then we have coordinates but that should

13
00:00:54,170 --> 00:01:00,200
then be an array of two numbers instead of an object of latitude and longitude which also are strings

14
00:01:00,230 --> 00:01:01,630
and not numbers.

15
00:01:01,670 --> 00:01:06,920
So this will again be quite a complex transformation but one which makes sense here and which we will

16
00:01:06,920 --> 00:01:07,640
need.

17
00:01:08,030 --> 00:01:15,770
I also want to include my email in the output and I want to include my date of birth data but I also

18
00:01:15,770 --> 00:01:22,310
want to transform this to basically flatten it to not have a nested document but simply have a birth

19
00:01:22,310 --> 00:01:27,020
date which is this and then an age field as a top level field.

20
00:01:27,020 --> 00:01:32,030
Now definitely feel free to pause and do this on your own but we will need a couple of operators we

21
00:01:32,150 --> 00:01:33,430
haven't touched on yet,

22
00:01:33,440 --> 00:01:38,210
so also feel free to follow along with me depending on whether you feel like diving into the official

23
00:01:38,210 --> 00:01:39,340
docs or not.

24
00:01:41,030 --> 00:01:44,940
So let's start with the location,

25
00:01:44,960 --> 00:01:50,840
now what I will do to keep this a bit more manageable, I will add an extra project stage and this already

26
00:01:50,840 --> 00:01:52,920
also shows you something else which is important,

27
00:01:53,030 --> 00:01:58,010
you can have the same stage multiple times and that's not that uncommon

28
00:01:58,010 --> 00:02:04,910
to be honest, often you also do some matching, sorting, grouping and then you project and then maybe you

29
00:02:04,910 --> 00:02:05,310
sort

30
00:02:05,330 --> 00:02:11,460
again, so often you have some in-between stages, here I split it across two stages to make it easier to read.

31
00:02:12,080 --> 00:02:12,690
So I will

32
00:02:12,710 --> 00:02:18,410
exclude my ID here already and therefore, I don't have to do it here because remember, this stage will

33
00:02:18,410 --> 00:02:19,980
receive the output of this stage,

34
00:02:20,090 --> 00:02:24,020
so if I exclude the ID here, it will not be available in the next stage,

35
00:02:24,260 --> 00:02:31,190
I should put this between curly braces though. Now I will add the name here,

36
00:02:31,700 --> 00:02:35,150
that should be included because the name is something I need in the next stage,

37
00:02:35,180 --> 00:02:38,390
so we have to make sure that we pass this embedded document

38
00:02:38,990 --> 00:02:41,810
and then I want to work on the location and so on.

39
00:02:41,810 --> 00:02:46,820
Now first of all, I'll pass on the e-mail because that will also just be in the next stage without any

40
00:02:46,820 --> 00:02:53,570
changes and therefore in the next stage, I'll also keep that email and well output it as a result of

41
00:02:53,570 --> 00:02:53,860
this

42
00:02:53,870 --> 00:02:54,920
stage too

43
00:02:55,160 --> 00:02:58,340
and now let's work on the location.

44
00:02:58,620 --> 00:03:04,100
Now the location at the moment is an embedded document with a lot of data that I'm currently not interested

45
00:03:04,100 --> 00:03:09,980
in for this operation and where I also well have it in the wrong format,

46
00:03:10,160 --> 00:03:18,060
I want to have a type and then coordinates which should be an array. So location should be a document,

47
00:03:18,110 --> 00:03:19,340
this is the case

48
00:03:19,520 --> 00:03:25,860
and in that document, I want to have a type field which will be point. And that is something we can do,

49
00:03:25,910 --> 00:03:29,060
we can add hardcoded values like this.

50
00:03:29,090 --> 00:03:30,390
So now I added this,

51
00:03:30,500 --> 00:03:38,990
now I also need to add my coordinates, the coordinates right now are a nested document like this,

52
00:03:38,990 --> 00:03:45,750
I need an array of coordinates though. And this is relatively simple to achieve,

53
00:03:45,830 --> 00:03:48,150
I simply set coordinates to an array

54
00:03:48,520 --> 00:03:53,810
and now the first value here will be dollar sign

55
00:03:53,810 --> 00:03:56,940
and now I need to access location coordinates

56
00:03:56,960 --> 00:04:02,330
longitude because remember, for a geo json object, the first value has to be a longitude.

57
00:04:02,750 --> 00:04:09,630
So I access location coordinates longitude here,

58
00:04:09,680 --> 00:04:13,160
so this field is getting accessed here,

59
00:04:13,280 --> 00:04:18,110
the dollar sign is important to tell mongodb that does this not a hardcoded string value but

60
00:04:18,110 --> 00:04:22,690
instead, the value stored in this path here in our document

61
00:04:23,030 --> 00:04:27,670
and the second element of the coordinates array is almost the same

62
00:04:27,680 --> 00:04:31,150
but of course here, I refer to latitude,

63
00:04:31,160 --> 00:04:34,930
so now this is my project phase as it is currently written.

64
00:04:35,300 --> 00:04:39,640
So now location should be such a geo json object,

65
00:04:39,650 --> 00:04:41,860
let's give this a try. For this,

66
00:04:41,870 --> 00:04:50,000
let's also include location which is this location because we get the output of the last stage as an input

67
00:04:50,000 --> 00:04:50,740
to this stage,

68
00:04:50,870 --> 00:04:56,180
so we will have that location key which is the result of this location transformation and let's include

69
00:04:56,180 --> 00:04:58,410
that in our new output.

70
00:04:58,490 --> 00:05:04,130
Now again, aggregate has to go on this line and pretty has to go on this line

71
00:05:04,130 --> 00:05:07,680
and now let's copy that entire stage here

72
00:05:09,640 --> 00:05:16,480
and let's run it. And what we can see is that for every person, we indeed have the full name,

73
00:05:16,560 --> 00:05:24,300
we have the email and location indeed is a point or is a geo json object with type point and coordinates.

74
00:05:24,300 --> 00:05:26,390
Now one thing you might also see is that

75
00:05:26,490 --> 00:05:32,340
this however is a string and not a number which it should be,

76
00:05:32,340 --> 00:05:39,750
now to change that, we have to convert this string. Converting data is something you often have to do,

77
00:05:40,010 --> 00:05:43,220
we'll have to do it later again for the birth date

78
00:05:43,230 --> 00:05:47,320
but now let's do it for our coordinates here.

79
00:05:47,460 --> 00:05:53,420
Of course mongodb also has a convert operator we can use,

80
00:05:53,500 --> 00:05:57,000
so we want to convert our coordinates here.

81
00:05:57,460 --> 00:06:03,870
So the value I store as longitude will be the result of an expression using the $convert

82
00:06:04,170 --> 00:06:05,830
operator, the

83
00:06:05,830 --> 00:06:11,830
$convert operator takes a document where you configure it, so where you configure its inputs and so on.

84
00:06:11,860 --> 00:06:14,280
Now convert needs a couple of fields,

85
00:06:14,410 --> 00:06:20,630
the first field is input and there you simply define which value should be converted,

86
00:06:20,680 --> 00:06:25,370
this of course is the value which is stored in a longitude field in our case.

87
00:06:25,420 --> 00:06:26,360
The second field

88
00:06:26,380 --> 00:06:33,910
and this is the last field that you need to pass is the to field, the to field defines the type you want to

89
00:06:33,910 --> 00:06:35,320
convert to.

90
00:06:35,470 --> 00:06:38,040
Now which types are available?

91
00:06:38,170 --> 00:06:43,560
This is of course also something you find in the official docs of the convert method, there

92
00:06:43,660 --> 00:06:52,510
you find that the to field takes these string identifiers to convert it to a double or to a decimal

93
00:06:52,720 --> 00:06:58,130
which is simply a constant with higher precision than a normal double and so on.

94
00:06:58,150 --> 00:07:02,150
You also can define on error and on null values,

95
00:07:02,200 --> 00:07:07,520
so default values that should be returned in case the transformation fails.

96
00:07:07,540 --> 00:07:15,010
So here I want to convert this to a double which is a floating point value and on error will be 0,

97
00:07:15,460 --> 00:07:19,330
on null will also be 0 or a 0.0,

98
00:07:19,330 --> 00:07:22,040
so we store or we return just 0

99
00:07:22,120 --> 00:07:30,530
if we fail to convert this. Now I can take this entire step here and replace my latitude step with it,

100
00:07:30,530 --> 00:07:35,240
of course replacing longitude with latitude down there again.

101
00:07:35,240 --> 00:07:38,460
So now these coordinates should be transformed,

102
00:07:38,480 --> 00:07:42,980
so let's now bring aggregate and pretty back into line

103
00:07:44,870 --> 00:07:50,630
and let's copy the entire stage or the entire pipeline and execute it.

104
00:07:50,640 --> 00:07:53,420
And now we see that the coordinates indeed are numbers,

105
00:07:53,550 --> 00:07:56,700
they are no longer wrapped in quotation marks.