1
00:00:02,380 --> 00:00:07,670
Now with all the projections we did thus far, let's now add another stage again

2
00:00:07,930 --> 00:00:10,120
after all our projections,

3
00:00:10,210 --> 00:00:18,490
so here I'll add another group stage because this is another common scenario that you first want to restructure

4
00:00:18,490 --> 00:00:19,070
your data,

5
00:00:19,090 --> 00:00:21,960
recalculate some fields as we are doing it,

6
00:00:22,030 --> 00:00:25,300
we're adding brand new fields which didn't exist before

7
00:00:25,600 --> 00:00:30,220
and then you want to group on these derived fields.

8
00:00:30,220 --> 00:00:37,240
In my case here I want to group on the birth date and yes, this was available before but not as a date

9
00:00:37,270 --> 00:00:37,730
at least,

10
00:00:37,750 --> 00:00:41,250
we converted it and I will now take advantage of this conversion.

11
00:00:41,290 --> 00:00:46,810
So I'll take advantage of my birth date which is now stored as a date in the group stage here.

12
00:00:46,960 --> 00:00:52,570
We saw the group stage before, we add _id here and then we define by which value we want to

13
00:00:52,570 --> 00:00:53,550
group.

14
00:00:53,590 --> 00:01:00,550
Now here I want to group by birthdate but not by the date but let's say by the year, so birth year is by

15
00:01:00,550 --> 00:01:02,250
which I want to group,

16
00:01:02,470 --> 00:01:09,160
therefore we add a document here and then the operator is the ISO year operator which basically retrieves

17
00:01:09,400 --> 00:01:12,200
the year out of a date.

18
00:01:12,370 --> 00:01:19,010
So the date then is stored in $birthdate pointing at my birth date field of course

19
00:01:19,120 --> 00:01:23,530
and now what do I want to do when grouping these documents?

20
00:01:23,990 --> 00:01:28,750
Well I want to calculate how many persons were born in this year,

21
00:01:28,750 --> 00:01:35,920
so num persons is what I calculate here and I use sum one to basically add one for every document that

22
00:01:35,920 --> 00:01:36,950
goes into that group

23
00:01:37,000 --> 00:01:42,430
and since one document represents one person, I will have the number of persons per birth year here at the

24
00:01:42,430 --> 00:01:42,950
end.

25
00:01:43,240 --> 00:01:49,300
So now with that, whoops this is ISO week year by the way, should be week year.

26
00:01:49,390 --> 00:01:58,030
So now let's copy that, paste it in here and you see, I got my years extracted here as a long number,

27
00:01:58,060 --> 00:02:00,200
this is what mongodb did for me

28
00:02:00,460 --> 00:02:03,150
and I see how many persons were born in there.

29
00:02:03,460 --> 00:02:09,160
Obviously my other projection phases are now rendered a bit useless since I don't take advantage of all the

30
00:02:09,320 --> 00:02:10,440
other fields I added here.

31
00:02:10,450 --> 00:02:11,790
This is just temporary,

32
00:02:11,800 --> 00:02:16,030
I just wanted to show you that you can group on the results of projections.

33
00:02:16,030 --> 00:02:17,890
So here I have the group stage,

34
00:02:17,950 --> 00:02:29,630
let's now maybe even sort and let's sort on the num persons in descending order to start with the most

35
00:02:29,630 --> 00:02:32,900
persons first. If I now enter this here,

36
00:02:32,900 --> 00:02:38,280
now we see most persons in the example dataset were born in 1955.