WEBVTT

1
00:00:00.000 --> 00:00:02.050
<v Instructor>Quick theory lecture on S3 Select</v>

2
00:00:02.050 --> 00:00:03.100
and Glacier Select.

3
00:00:03.100 --> 00:00:05.640
The the idea is we want to retrieve less data,

4
00:00:05.640 --> 00:00:08.710
so subsets of what we're requesting using SQL

5
00:00:08.710 --> 00:00:10.970
by performing server side filtering,

6
00:00:10.970 --> 00:00:12.930
and so the SQLs queries are quite simple.

7
00:00:12.930 --> 00:00:15.280
They can only be used to filter by rows and columns

8
00:00:15.280 --> 00:00:16.980
so they're very simple SQL statements.

9
00:00:16.980 --> 00:00:19.380
You cannot do aggregations or anything like this,

10
00:00:19.380 --> 00:00:23.100
and you will use less network and less CPU cost client-side

11
00:00:23.100 --> 00:00:24.930
because you don't retrieve the full file.

12
00:00:24.930 --> 00:00:28.070
S3 will perform the select, the filtering for you

13
00:00:28.070 --> 00:00:30.140
and only return to you what you need.

14
00:00:30.140 --> 00:00:32.930
So the idea is that before you have Amazon S3

15
00:00:32.930 --> 00:00:35.360
sending all the data into your application

16
00:00:35.360 --> 00:00:37.110
and then you have to filter it application-side

17
00:00:37.110 --> 00:00:38.400
to find the right rows you want

18
00:00:38.400 --> 00:00:40.250
and only keep the columns you want,

19
00:00:40.250 --> 00:00:44.540
and after you request the data from S3 using S3 Select

20
00:00:44.540 --> 00:00:46.800
and it only gives you the data you need,

21
00:00:46.800 --> 00:00:48.810
the columns you want and the rows you want

22
00:00:48.810 --> 00:00:51.030
and the results Amazon is telling you is

23
00:00:51.030 --> 00:00:54.530
that you are up to 400% faster and up to 80% cheaper

24
00:00:54.530 --> 00:00:56.600
because you have less network traffic going through

25
00:00:56.600 --> 00:00:59.740
and the filtering happens server-side, okay.

26
00:00:59.740 --> 00:01:02.410
So similarly, let's just do another diagram.

27
00:01:02.410 --> 00:01:04.490
We have the client asking to get a CSV file

28
00:01:04.490 --> 00:01:07.890
with S3 Select to only get a few columns and a few rows.

29
00:01:07.890 --> 00:01:09.980
Amazon S3 will perform server-side filtering

30
00:01:09.980 --> 00:01:12.370
on that CSV file to find the right columns

31
00:01:12.370 --> 00:01:13.450
and the rows we want

32
00:01:13.450 --> 00:01:16.870
and send back the data filtered back to our client,

33
00:01:16.870 --> 00:01:20.370
so obviously, less network, less CPU, and faster,

34
00:01:20.370 --> 00:01:22.050
so this is great.

35
00:01:22.050 --> 00:01:23.710
To summarize from an exam perspective,

36
00:01:23.710 --> 00:01:26.340
any time you see filtering of data server-side

37
00:01:26.340 --> 00:01:31.140
in S3 to get less, think about S3 Select English or select,

38
00:01:31.140 --> 00:01:32.540
that works on Glacier as well

39
00:01:32.540 --> 00:01:34.530
and then for more complex querying,

40
00:01:34.530 --> 00:01:36.290
that's gonna be server less on S3,

41
00:01:36.290 --> 00:01:37.330
you'll see in the future lectures

42
00:01:37.330 --> 00:01:39.390
we have something called Amazon Athena.

43
00:01:39.390 --> 00:01:40.223
All right.

44
00:01:40.223 --> 00:01:42.220
That's it. I will see you in the next lecture.

