1
00:00:06,810 --> 00:00:08,520
- Another common element in a script

2
00:00:08,520 --> 00:00:10,710
is the pattern-matching operator.

3
00:00:10,710 --> 00:00:11,970
The pattern-matching operator

4
00:00:11,970 --> 00:00:14,340
is all about cleaning up strings.

5
00:00:14,340 --> 00:00:16,470
So if a user provides information,

6
00:00:16,470 --> 00:00:18,390
but there's too much information

7
00:00:18,390 --> 00:00:20,430
and you only need to have a part of it,

8
00:00:20,430 --> 00:00:23,430
then pattern-matching operator
is what you are going to use.

9
00:00:24,480 --> 00:00:25,920
So how does it work?

10
00:00:25,920 --> 00:00:27,510
Well, in a pattern-matching operator,

11
00:00:27,510 --> 00:00:30,210
you perform an operation on a variable.

12
00:00:30,210 --> 00:00:32,010
If the variable is $1,

13
00:00:32,010 --> 00:00:34,080
which is the first positional parameter

14
00:00:34,080 --> 00:00:35,970
or the first argument that was provided

15
00:00:35,970 --> 00:00:37,440
while running the script,

16
00:00:37,440 --> 00:00:41,430
you can put the $1 between curly braces,

17
00:00:41,430 --> 00:00:44,256
and do ${1#}.

18
00:00:44,256 --> 00:00:46,830
That's not really a
pattern-matching operator,

19
00:00:46,830 --> 00:00:49,680
but it prints the length of the string.

20
00:00:49,680 --> 00:00:51,033
And that can be useful.

21
00:00:51,930 --> 00:00:56,550
Now if you would use ${1#string},

22
00:00:56,550 --> 00:00:58,140
then you are going to search for "string"

23
00:00:58,140 --> 00:01:00,180
and you'll remove the
shortest match of the string

24
00:01:00,180 --> 00:01:02,610
from the front end of $1.

25
00:01:02,610 --> 00:01:03,960
There's also double-hash,

26
00:01:03,960 --> 00:01:06,810
and that will remove the
longest match of the string,

27
00:01:06,810 --> 00:01:08,163
also from the front end.

28
00:01:09,720 --> 00:01:12,060
Then we have %string,

29
00:01:12,060 --> 00:01:14,550
which removes the shortest
match of the string,

30
00:01:14,550 --> 00:01:16,020
and %%string,

31
00:01:16,020 --> 00:01:17,970
which removes the longest
match of the string

32
00:01:17,970 --> 00:01:19,183
from the back end of $1.

33
00:01:20,880 --> 00:01:22,410
Let's have a look at some examples

34
00:01:22,410 --> 00:01:26,253
in patterns.sh, which is in
the course Git repository.

35
00:01:31,260 --> 00:01:33,450
Okay, let's have a look at patterns.sh,

36
00:01:33,450 --> 00:01:36,270
which is a simple example script

37
00:01:36,270 --> 00:01:38,940
which is demonstrating
what this is all about.

38
00:01:38,940 --> 00:01:40,470
So in this example script,

39
00:01:40,470 --> 00:01:43,590
we are going to provide
a string as argument

40
00:01:43,590 --> 00:01:44,970
while running the script,

41
00:01:44,970 --> 00:01:47,940
like /usr/bin/blah,

42
00:01:47,940 --> 00:01:50,130
but you can use anything you want.

43
00:01:50,130 --> 00:01:51,060
This script is really

44
00:01:51,060 --> 00:01:52,230
a script that is looking for filenames,

45
00:01:52,230 --> 00:01:55,110
so it should look like a filename,

46
00:01:55,110 --> 00:01:57,540
that's the only thing that matters.

47
00:01:57,540 --> 00:01:58,860
Now, what are we going to do?

48
00:01:58,860 --> 00:02:01,830
Well, first we set the variable "filename"

49
00:02:01,830 --> 00:02:03,793
and filename is set to $1,

50
00:02:04,920 --> 00:02:07,590
and then we have double-hash */,

51
00:02:08,640 --> 00:02:11,790
So we are looking for */,

52
00:02:11,790 --> 00:02:14,313
which is a / with anything in front of it.

53
00:02:15,720 --> 00:02:19,590
Then we are looking for the
longest match of this pattern.

54
00:02:19,590 --> 00:02:20,880
What is the longest match?

55
00:02:20,880 --> 00:02:24,030
Well, if we look from left
to right in the string,

56
00:02:24,030 --> 00:02:26,460
now we are really searching
for the last slash,

57
00:02:26,460 --> 00:02:29,430
and we remove it with all
that is in front of it,

58
00:02:29,430 --> 00:02:31,740
so that the filename is remaining.

59
00:02:31,740 --> 00:02:36,740
And then we use echo filename
is '${1 hash-hash */}'

60
00:02:37,680 --> 00:02:39,450
between single quotes.

61
00:02:39,450 --> 00:02:40,530
The single quotes make

62
00:02:40,530 --> 00:02:43,050
that we escape the
pattern-matching operator,

63
00:02:43,050 --> 00:02:46,140
and we just print what is going to happen.

64
00:02:46,140 --> 00:02:51,120
That's easy for understanding
what is going to happen.

65
00:02:51,120 --> 00:02:54,150
And in the next line, the
name of the file is $filename.

66
00:02:54,150 --> 00:02:55,950
That is printing the "filename" variable

67
00:02:55,950 --> 00:02:59,460
that was set according to the
pattern-matching operator.

68
00:02:59,460 --> 00:03:01,830
We do something similar
for the directory name,

69
00:03:01,830 --> 00:03:06,390
but this time the
pattern-matching operator is %/*.

70
00:03:06,390 --> 00:03:07,470
What does that mean?

71
00:03:07,470 --> 00:03:09,180
That means that from right to left

72
00:03:09,180 --> 00:03:14,100
we are searching for the
pattern / followed by anything.

73
00:03:14,100 --> 00:03:18,030
It's a single %, so that
will be shortest match.

74
00:03:18,030 --> 00:03:21,540
And the shortest match in
this example would be /blah.

75
00:03:21,540 --> 00:03:25,140
So the remaining parts is
the directory name, /usr/bin,

76
00:03:25,140 --> 00:03:27,480
and that is what's going to be printed.

77
00:03:27,480 --> 00:03:30,433
Let's run it so that we can
see what's going to happen.

78
00:03:30,433 --> 00:03:33,433
(keyboard clacking)

79
00:03:34,590 --> 00:03:36,540
So in the script it is important

80
00:03:36,540 --> 00:03:40,473
that we provide an
argument, /usr/bin/whatever.

81
00:03:41,880 --> 00:03:43,950
Really, it's not really important

82
00:03:43,950 --> 00:03:46,170
if it's an existing file or anything,

83
00:03:46,170 --> 00:03:48,450
it's just about what the
pattern-matching operator

84
00:03:48,450 --> 00:03:49,470
is going to do.

85
00:03:49,470 --> 00:03:50,760
And what do we see?

86
00:03:50,760 --> 00:03:52,890
We see that it keeps
the filename "whatever"

87
00:03:52,890 --> 00:03:56,820
and the directory name,
which is set to /usr/bin.

88
00:03:56,820 --> 00:03:58,830
Now we are going to do
something interesting,

89
00:03:58,830 --> 00:04:01,650
and we apply a couple of modifications.

90
00:04:01,650 --> 00:04:04,830
So for filename I am going to make that

91
00:04:04,830 --> 00:04:08,730
no longer hash-hash, but single-hash */.

92
00:04:10,620 --> 00:04:14,043
And for directory name, I
make that double-percent.

93
00:04:16,440 --> 00:04:20,460
So we will be going in the
opposite direction here,

94
00:04:20,460 --> 00:04:23,400
and then we are going to
check what is going to happen.

95
00:04:23,400 --> 00:04:24,510
I would like to invite you

96
00:04:24,510 --> 00:04:26,700
to try to understand
what is going to happen

97
00:04:26,700 --> 00:04:27,783
before running it.

98
00:04:28,740 --> 00:04:31,980
And once you have an
idea, then you run it.

99
00:04:31,980 --> 00:04:35,700
To give you a hint,
single-hash is shortest match.

100
00:04:35,700 --> 00:04:40,700
And to give you another hint,
* is zero or more characters.

101
00:04:40,710 --> 00:04:42,480
So we are looking for the shortest match

102
00:04:42,480 --> 00:04:46,653
of a / with zero or more
characters in front of it.

103
00:04:47,820 --> 00:04:48,813
Let's check it out.

104
00:04:52,590 --> 00:04:57,420
And there you can see what
the single-hash */ is doing.

105
00:04:57,420 --> 00:05:00,000
It's just removing the
slash in the beginning.

106
00:05:00,000 --> 00:05:02,940
So the remaining part is usr/bin/whatever.

107
00:05:02,940 --> 00:05:04,650
And for the directory name,

108
00:05:04,650 --> 00:05:05,700
well, the directory name

109
00:05:05,700 --> 00:05:08,340
is looking for the longest match

110
00:05:08,340 --> 00:05:11,280
of the pattern / followed by anything,

111
00:05:11,280 --> 00:05:13,860
and that happens to be the entire pattern.

112
00:05:13,860 --> 00:05:17,130
So the name of the
directory is nothing at all.

113
00:05:17,130 --> 00:05:20,250
And that's how you can use
pattern-matching operators.

114
00:05:20,250 --> 00:05:23,370
Pattern-matching operators
are clean and they are fast.

115
00:05:23,370 --> 00:05:25,590
They are best internal solution,

116
00:05:25,590 --> 00:05:28,260
so you don't require
anything to be loaded,

117
00:05:28,260 --> 00:05:30,570
and that is why you see
pattern-matching operators

118
00:05:30,570 --> 00:05:31,983
quite often in scripts.