1
00:00:00,000 --> 00:00:05,250
All right, you know how to slice lists.

2
00:00:05,275 --> 00:00:07,296
[No audio]

3
00:00:07,320 --> 00:00:10,140
0 to 1 will give you the first item

4
00:00:10,140 --> 00:00:12,540
of the list, and 0 to 2, the first

5
00:00:12,540 --> 00:00:14,488
and the second item of the list, and so on.

6
00:00:15,178 --> 00:00:17,585
With NumPy arrays, you do basically

7
00:00:17,610 --> 00:00:20,820
the same thing. Except with NumPy

8
00:00:20,820 --> 00:00:22,860
arrays, sometimes you have 2 or more

9
00:00:22,860 --> 00:00:26,250
dimensionals, or 3. So let me go

10
00:00:26,250 --> 00:00:29,310
ahead and I'll create, I'll start with a

11
00:00:29,310 --> 00:00:32,820
2 dimensional array. So I'll get that

12
00:00:32,820 --> 00:00:36,570
from these variable. So this is the NumPy

13
00:00:36,570 --> 00:00:39,300
array that we'll be working with, and

14
00:00:39,300 --> 00:00:44,340
yeah, if you want to extract this number

15
00:00:44,340 --> 00:00:47,760
here, this here, this one, and this, so

16
00:00:47,760 --> 00:00:51,540
these four numbers, well, you'd first needs

17
00:00:51,540 --> 00:00:54,420
to set the index of the rows that you

18
00:00:54,420 --> 00:01:01,080
want to slice. So that'd be 0 to 2, which

19
00:01:01,080 --> 00:01:03,720
gives you the first 2 rows. Actually,

20
00:01:03,720 --> 00:01:06,450
we can try that, and you'll see that you

21
00:01:06,450 --> 00:01:09,540
get the first 2 rows. Now, if you want

22
00:01:09,755 --> 00:01:12,275
only this portion here, so these

23
00:01:12,300 --> 00:01:14,970
two numbers and these two, you would

24
00:01:14,970 --> 00:01:16,030
want to pass

25
00:01:16,055 --> 00:01:18,055
[No audio]

26
00:01:18,080 --> 00:01:20,460
comma there, and now the index

27
00:01:20,460 --> 00:01:22,980
for the columns, which would be, you

28
00:01:22,980 --> 00:01:27,187
know, 0 here, one, so two, three,

29
00:01:27,211 --> 00:01:30,480
and yeah 2 to 4 maybe, you know, that

30
00:01:30,480 --> 00:01:32,688
should do it, and that's it.

31
00:01:32,712 --> 00:01:35,081
[No audio]

32
00:01:35,106 --> 00:01:37,770
That means you have this indexing system. So it

33
00:01:37,770 --> 00:01:40,590
start from 0 for rows, and then row, the

34
00:01:40,590 --> 00:01:43,380
second row is 1, and 2, and so on,

35
00:01:43,500 --> 00:01:47,970
for columns are the same thing, 0, 1, 2, 3,

36
00:01:48,090 --> 00:01:51,720
and 4, and actually, if you like, you

37
00:01:51,745 --> 00:01:55,045
can see the shape of your NumPy array.

38
00:01:55,380 --> 00:01:57,780
So it's 3 by 5, which means 0

39
00:01:57,780 --> 00:02:02,220
to 2, and 0 to 4, great, and of

40
00:02:02,220 --> 00:02:05,160
course, you can use the convention of

41
00:02:05,190 --> 00:02:08,070
the list indexing. So you can pass 0 and

42
00:02:08,070 --> 00:02:10,410
everything after that or just like that.

43
00:02:11,850 --> 00:02:15,270
So you get all the rows and only columns

44
00:02:15,270 --> 00:02:18,390
from 2 to 3. Similarly, you can

45
00:02:18,390 --> 00:02:19,915
get only one value if you like.

46
00:02:19,939 --> 00:02:24,169
[No audio]

47
00:02:24,194 --> 00:02:27,600
Not particularly this one, because 3 is

48
00:02:27,600 --> 00:02:31,650
out of bounds for axis 0. So axis 0

49
00:02:31,710 --> 00:02:35,010
is the horizontal axis, which says that

50
00:02:35,010 --> 00:02:38,940
there is no row with index 3. So we

51
00:02:38,940 --> 00:02:43,410
set is 0, 1, 2. So let's pass 2 there, and you

52
00:02:43,410 --> 00:02:47,790
get 182, which is the very last item

53
00:02:47,790 --> 00:02:50,400
of the NumPy array. And yeah, that's about

54
00:02:50,430 --> 00:02:52,710
indexing. How about iterating through a

55
00:02:52,710 --> 00:02:55,920
NumPy array? Well, there are two ways to

56
00:02:55,920 --> 00:02:58,290
do that. The first way is to say, let's

57
00:02:58,315 --> 00:03:02,635
say for i in your numpy array,

58
00:03:04,080 --> 00:03:09,120
print i. And what that will do is it will

59
00:03:09,120 --> 00:03:12,720
print out the rows of your NumPy

60
00:03:12,720 --> 00:03:16,650
arrays, and so what these i access is in

61
00:03:16,650 --> 00:03:19,350
your NumPy array is the rows. So in each

62
00:03:19,350 --> 00:03:21,570
iteration, it gets the first row and it

63
00:03:21,570 --> 00:03:23,910
prints it out in the first iteration,

64
00:03:23,910 --> 00:03:25,922
then the second one, and then the third one,

65
00:03:27,629 --> 00:03:30,180
and if you want to iterate through

66
00:03:30,180 --> 00:03:32,670
columns, you'd want to access the

67
00:03:32,700 --> 00:03:36,660
transposed version of your NumPy array, and

68
00:03:36,660 --> 00:03:39,270
that's how you do it, and lastly, if

69
00:03:39,270 --> 00:03:41,460
you want to iterate value by value,

70
00:03:41,940 --> 00:03:49,020
you'd say for i in image and .flat

71
00:03:49,044 --> 00:03:51,725
[No audio]

72
00:03:51,750 --> 00:03:53,355
print i.

73
00:03:53,379 --> 00:03:55,616
[No audio]

74
00:03:55,641 --> 00:03:56,910
So this properly what it

75
00:03:56,910 --> 00:03:59,520
does is it allows you to access the

76
00:03:59,580 --> 00:04:02,100
values of your NumPy array one by one.

77
00:04:03,425 --> 00:04:05,345
And that's about indexing, slicing, and

78
00:04:05,370 --> 00:04:07,866
iterating. I'll see you in the next lecture.