1
00:00:02,360 --> 00:00:06,510
Now that we had a look at a one-to-many relationship where embedding made sense,

2
00:00:06,680 --> 00:00:12,530
let's have a look at a relationship, a one-to-many relationship where splitting it in collections might

3
00:00:12,530 --> 00:00:14,740
make sense. Here

4
00:00:14,900 --> 00:00:22,460
I got a couple of cities like New York, Berlin and so on and I got the citizens, let's say we have a database,

5
00:00:22,460 --> 00:00:30,170
a collection which holds all the major cities of the world and then a list of every person living in

6
00:00:30,170 --> 00:00:35,910
that city, like the registry of the state where all people are registered, let's say.

7
00:00:36,200 --> 00:00:37,960
So we get something like this,

8
00:00:37,970 --> 00:00:39,700
the problem we'll face here,

9
00:00:39,800 --> 00:00:45,160
theoretically that would make sense to be a one-to-many relationship where you embed your documents

10
00:00:45,350 --> 00:00:53,030
but for one from application perspective, you might often be interested in fetching all the cities with

11
00:00:53,030 --> 00:00:58,220
their metadata but you don't want to fetch all the citizens along with the cities because in New York

12
00:00:58,220 --> 00:01:04,040
where we have like millions of people, that would be a lot of overhead data which we don't need which

13
00:01:04,040 --> 00:01:08,760
takes very long to fetch and which is a lot of data to transfer over the wire.

14
00:01:09,050 --> 00:01:15,590
And additionally for big cities with millions of citizens, you might indeed hit this 16mb limit

15
00:01:15,590 --> 00:01:17,550
for the overall document

16
00:01:17,630 --> 00:01:21,720
and keep in mind, that limit counts towards a document in nested documents,

17
00:01:21,740 --> 00:01:27,800
so it's not 16mb per nested element, it's 16mb for the overall document,

18
00:01:27,800 --> 00:01:32,130
so you might face this limit too or might hit this limit too.

19
00:01:32,330 --> 00:01:35,440
So for technical and application reasons, here

20
00:01:35,480 --> 00:01:37,830
splitting it up might make sense.

21
00:01:38,120 --> 00:01:44,990
So here, you might have your city data database and in there, you have your cities

22
00:01:47,860 --> 00:01:58,570
collection where you insert a city like the name, New York City coordinates which is a nested document

23
00:01:58,870 --> 00:02:00,150
with a latitude

24
00:02:00,190 --> 00:02:01,540
and this is arbitrary,

25
00:02:01,540 --> 00:02:07,310
this is not a real latitude and longitude, something like this

26
00:02:07,450 --> 00:02:09,810
and you store no data about the citizens in there,

27
00:02:09,850 --> 00:02:17,290
so you also have no citizens array in here with all the IDs of the citizens because that would be a super

28
00:02:17,290 --> 00:02:19,570
long array too if you have millions of citizens.

29
00:02:19,570 --> 00:02:21,460
So you just have the metadata

30
00:02:25,990 --> 00:02:26,790
and there

31
00:02:26,800 --> 00:02:29,190
I got to access square bracket.

32
00:02:29,410 --> 00:02:30,280
So you just have

33
00:02:30,280 --> 00:02:31,050
that stored 

34
00:02:31,120 --> 00:02:36,170
and therefore if you find that one city, it looks like this.

35
00:02:36,170 --> 00:02:40,450
And you might also store the citizens but you will do this,

36
00:02:40,450 --> 00:02:47,500
citizens in a separate collection and then there you might have your different citizens with the name

37
00:02:47,720 --> 00:02:52,760
Max Schwarzmuller and the city ID maybe

38
00:02:52,890 --> 00:02:58,710
and that ID could be in New York City to match it by the name or you simply use that object ID of

39
00:02:58,710 --> 00:03:03,120
course, you want to use something which is unique so let's go for the objectId.

40
00:03:03,360 --> 00:03:08,730
So that would be one citizen and then we have another citizen where you also have a name,

41
00:03:11,740 --> 00:03:13,120
Manuel Lorenz and

42
00:03:13,190 --> 00:03:14,960
we're not really living in New York

43
00:03:14,980 --> 00:03:18,470
but just as an example here, just store that like this.

44
00:03:18,490 --> 00:03:23,710
Now insertMany needs square brackets around all the objects you're inserting, so don't forget to

45
00:03:23,710 --> 00:03:25,970
add these, hit enter

46
00:03:26,350 --> 00:03:32,030
and I got a typo, I was missing the closing parentheses after objectId.

47
00:03:32,520 --> 00:03:34,300
And now I inserted my citizens

48
00:03:34,570 --> 00:03:40,530
and if I inspect my citizens, if I find them all and I pretty print this,

49
00:03:40,600 --> 00:03:41,740
here are my citizens

50
00:03:41,800 --> 00:03:46,840
and now I can always retrieve just as citizens and match them to their city with the city ID if I need

51
00:03:46,840 --> 00:03:53,920
to or fetch just the cities and I'll not hit the document limitation of 16mb for the city because

52
00:03:53,920 --> 00:04:00,210
I don't embed all my citizens and I also don't unnecessarily fetch all the citizens along with the city

53
00:04:00,280 --> 00:04:02,290
if I'm only interested in the city metadata.