1
00:00:06,720 --> 00:00:10,050
- In this video, you will learn
about regular expressions.

2
00:00:10,050 --> 00:00:12,180
So, regular expressions are text patterns

3
00:00:12,180 --> 00:00:14,670
that are used by tools
like grep and others.

4
00:00:14,670 --> 00:00:15,990
And why is it useful?

5
00:00:15,990 --> 00:00:17,280
Well that is useful in cases

6
00:00:17,280 --> 00:00:20,130
where you don't know exactly
what you are looking for.

7
00:00:20,130 --> 00:00:22,620
Now, you can specify the pattern,

8
00:00:22,620 --> 00:00:26,790
a little bit like wild cards,
in a shell environment.

9
00:00:26,790 --> 00:00:29,040
If you are going to use
regular expressions,

10
00:00:29,040 --> 00:00:31,530
you should always put
your regular expression,

11
00:00:31,530 --> 00:00:35,853
also shortly refer to as
"regex," between single quotes.

12
00:00:36,780 --> 00:00:38,670
That is because regular expressions

13
00:00:38,670 --> 00:00:40,620
are using characters that can

14
00:00:40,620 --> 00:00:42,540
also be interpreted by the shell.

15
00:00:42,540 --> 00:00:44,490
And you want to avoid any problems

16
00:00:44,490 --> 00:00:46,743
because of misinterpretation.

17
00:00:48,000 --> 00:00:51,090
That's also one of the big
things in regular expressions.

18
00:00:51,090 --> 00:00:52,260
The regular expressions

19
00:00:52,260 --> 00:00:55,590
sometimes looks like shell character

20
00:00:55,590 --> 00:00:57,000
but they are not.

21
00:00:57,000 --> 00:00:59,670
So, don't confuse the regular expressions,

22
00:00:59,670 --> 00:01:03,780
with globbing, which is
shell wildcards, and so on,

23
00:01:03,780 --> 00:01:06,570
because they look the
same but they are not.

24
00:01:06,570 --> 00:01:08,272
Look at this example for instance,

25
00:01:08,272 --> 00:01:13,272
grep single quote, a star,
single quote, on a star.

26
00:01:13,290 --> 00:01:15,900
So the single quotes are identifying

27
00:01:15,900 --> 00:01:18,720
that the first 'a star'
is a regular expression.

28
00:01:18,720 --> 00:01:23,340
And the second a star is shell globbing.

29
00:01:23,340 --> 00:01:26,010
Regular expressions are very powerful

30
00:01:26,010 --> 00:01:28,097
and you are going to appreciate them.

31
00:01:28,097 --> 00:01:30,690
But they only work in specific utilities

32
00:01:30,690 --> 00:01:32,970
like grep and vim and awk and sed.

33
00:01:32,970 --> 00:01:35,190
There's a couple more
utilities you will find

34
00:01:35,190 --> 00:01:36,453
your way automatically.

35
00:01:37,290 --> 00:01:38,370
What you should to remember

36
00:01:38,370 --> 00:01:41,460
If your utility is about finding text

37
00:01:41,460 --> 00:01:43,380
and about treating text,

38
00:01:43,380 --> 00:01:44,340
then there's a chance

39
00:01:44,340 --> 00:01:47,100
that it works with regular expressions.

40
00:01:47,100 --> 00:01:48,650
I will give you an introduction

41
00:01:48,650 --> 00:01:51,210
to the most useful regular expressions.

42
00:01:51,210 --> 00:01:53,460
For more detailed information

43
00:01:53,460 --> 00:01:57,150
consider reading "Man 7 regex",

44
00:01:57,150 --> 00:02:01,110
where the seven is referring
to man section seven,

45
00:02:01,110 --> 00:02:03,993
to get information about
the regular expressions.

46
00:02:05,160 --> 00:02:07,410
Now, regular expressions are powerful

47
00:02:07,410 --> 00:02:10,860
but regular expressions
also are confusing.

48
00:02:10,860 --> 00:02:12,000
Now, why is that?

49
00:02:12,000 --> 00:02:14,010
Well, that is because
regular expressions work

50
00:02:14,010 --> 00:02:18,630
with many tools, but
there is also this thing

51
00:02:18,630 --> 00:02:21,330
called extended regular expressions

52
00:02:21,330 --> 00:02:23,280
and extended regular expressions

53
00:02:23,280 --> 00:02:24,690
don't always work.

54
00:02:24,690 --> 00:02:27,270
And a regular expression
is either a normal

55
00:02:27,270 --> 00:02:29,460
or an extended regular expression.

56
00:02:29,460 --> 00:02:31,920
And if you use an extended
regular expression

57
00:02:31,920 --> 00:02:36,150
without using an extended
regular expression option

58
00:02:36,150 --> 00:02:38,100
then it's not going to work.

59
00:02:38,100 --> 00:02:39,510
Starting to feel confused?

60
00:02:39,510 --> 00:02:43,080
I can imagine, but I will
clarify once I demonstrate.

61
00:02:43,080 --> 00:02:46,380
You should also notice at
some scripting languages

62
00:02:46,380 --> 00:02:47,610
like Pearl for instance,

63
00:02:47,610 --> 00:02:49,743
come with their own regular expressions.

64
00:02:50,730 --> 00:02:52,860
And that makes that if
you're going to look

65
00:02:52,860 --> 00:02:55,442
up regular expressions
on Google, for instance,

66
00:02:55,442 --> 00:02:57,300
you will find information

67
00:02:57,300 --> 00:02:59,812
that doesn't match your environment

68
00:02:59,812 --> 00:03:02,133
and it won't work.

69
00:03:03,709 --> 00:03:07,260
Also confusing is that
there is so much information

70
00:03:07,260 --> 00:03:09,120
about regular expressions

71
00:03:09,120 --> 00:03:11,190
that it is difficult
to find the difference

72
00:03:11,190 --> 00:03:12,420
between the things that matter

73
00:03:12,420 --> 00:03:14,670
and the things that don't matter.

74
00:03:14,670 --> 00:03:17,790
Now, I've put together a slight

75
00:03:17,790 --> 00:03:21,390
that contains all of the most
common, regular expressions.

76
00:03:21,390 --> 00:03:22,223
And you know what?

77
00:03:22,223 --> 00:03:26,300
I am just going to put the entire slide on

78
00:03:27,240 --> 00:03:30,360
on the screen so that we can
talk about these examples

79
00:03:30,360 --> 00:03:35,360
and you can use the slide for
your own reference purposes.

80
00:03:41,598 --> 00:03:46,098
All right, let me start with
grep, grep Anna or users.

81
00:03:48,330 --> 00:03:49,260
What is that doing?

82
00:03:49,260 --> 00:03:53,313
This is showing all lines
that contain the text Anna.

83
00:03:54,180 --> 00:03:57,210
Notice that Anna in this
case is a regular expression

84
00:03:57,210 --> 00:04:00,090
in order to be interpreted
as a regular expression

85
00:04:00,090 --> 00:04:01,860
if you really want to do it a pure way

86
00:04:01,860 --> 00:04:03,960
put it between single quotes.

87
00:04:03,960 --> 00:04:06,060
No difference but it's good practice

88
00:04:06,060 --> 00:04:09,600
to identify your regular
expressions using single quotes.

89
00:04:09,600 --> 00:04:12,242
First leg X I want to introduce is carrot,

90
00:04:12,242 --> 00:04:16,713
which is looking for the pattern
at the beginning of a line.

91
00:04:17,730 --> 00:04:20,760
Likewise, we can also use dollar,

92
00:04:20,760 --> 00:04:22,170
which is doing the opposite,

93
00:04:22,170 --> 00:04:23,610
which is looking for the pattern

94
00:04:23,610 --> 00:04:24,990
at the end of the line,

95
00:04:24,990 --> 00:04:26,580
and you can combine them as well.

96
00:04:26,580 --> 00:04:29,160
So carrot, dollar is looking

97
00:04:29,160 --> 00:04:32,493
for lines starting and ending with.

98
00:04:33,462 --> 00:04:36,210
Now another one that is interesting

99
00:04:36,210 --> 00:04:38,970
is the slice B slash B

100
00:04:38,970 --> 00:04:40,890
is for end of word.

101
00:04:40,890 --> 00:04:43,320
Let's get back to the original start.

102
00:04:43,320 --> 00:04:44,640
Here we go.

103
00:04:44,640 --> 00:04:47,343
Look at the second line,
which is Annabelle.

104
00:04:48,210 --> 00:04:50,040
I want to do a regular expression

105
00:04:50,040 --> 00:04:52,350
that will work on words only.

106
00:04:52,350 --> 00:04:54,120
So how are we going to do that?

107
00:04:54,120 --> 00:04:58,830
Well, I am going to use grab Anna slash B

108
00:04:58,830 --> 00:05:02,370
slash B is the word marker on users.

109
00:05:02,370 --> 00:05:06,330
And there we can see
that it prints all lines

110
00:05:06,330 --> 00:05:07,803
that have Anna as a word,

111
00:05:08,670 --> 00:05:11,280
but not lines that contain it somewhere

112
00:05:11,280 --> 00:05:12,780
in another string.

113
00:05:12,780 --> 00:05:14,695
Now we have other regular expressions

114
00:05:14,695 --> 00:05:18,030
which are basically repeating operators

115
00:05:18,030 --> 00:05:20,970
and the repeating
operators are interesting,

116
00:05:20,970 --> 00:05:24,900
but I need to modify my
users file a little bit

117
00:05:24,900 --> 00:05:28,080
to show you how it works.

118
00:05:28,080 --> 00:05:33,080
So I am adding BT and bit
and bite and bolt and boot

119
00:05:36,360 --> 00:05:39,750
and boot and well, something like this.

120
00:05:39,750 --> 00:05:41,433
That should be good enough.

121
00:05:42,990 --> 00:05:46,560
So let's talk about
these repeating operators

122
00:05:46,560 --> 00:05:50,053
to start with I am going
to show you grep 'b'

123
00:05:53,193 --> 00:05:55,280
B I star T on users.

124
00:05:58,080 --> 00:05:59,400
Now this is an interesting one

125
00:05:59,400 --> 00:06:00,263
because here we have the star

126
00:06:00,263 --> 00:06:04,590
and the star means zero or more times.

127
00:06:04,590 --> 00:06:05,423
So what do we see?

128
00:06:05,423 --> 00:06:08,550
We see, we get match on
BT on bit and on bite.

129
00:06:08,550 --> 00:06:11,719
So we are looking for
zero or more occurrences

130
00:06:11,719 --> 00:06:14,733
of the, of the Is.

131
00:06:16,200 --> 00:06:19,440
You can also do that in
another way, by using a plus.

132
00:06:19,440 --> 00:06:23,133
A plus if you replace
the, the star with a plus

133
00:06:23,133 --> 00:06:26,443
then you are looking for one or more times

134
00:06:26,443 --> 00:06:28,650
but as you can see, that doesn't work.

135
00:06:28,650 --> 00:06:29,483
How come?

136
00:06:29,483 --> 00:06:30,840
Well, that is because this is one

137
00:06:30,840 --> 00:06:34,290
of these famous, extended
regular expressions.

138
00:06:34,290 --> 00:06:37,140
And if you are going to use
extended regular expressions

139
00:06:37,140 --> 00:06:37,973
with grep,

140
00:06:37,973 --> 00:06:39,793
you need grep minus uppercase E

141
00:06:39,793 --> 00:06:42,480
minus uppercase E will look

142
00:06:42,480 --> 00:06:45,030
for the extended regular expression

143
00:06:45,030 --> 00:06:46,770
and without minus uppercase E

144
00:06:46,770 --> 00:06:49,706
it is not interpreted the right way.

145
00:06:49,706 --> 00:06:51,780
Now we also have question mark

146
00:06:51,780 --> 00:06:52,740
what is question mark

147
00:06:52,740 --> 00:06:54,930
Question mark is zero or one time

148
00:06:54,930 --> 00:06:57,960
also an extended regular expression.

149
00:06:57,960 --> 00:07:02,960
So you want to get a match
on, on zero or one ice

150
00:07:04,770 --> 00:07:06,510
Well, this is what we get.

151
00:07:06,510 --> 00:07:08,370
That's not the most convincing example.

152
00:07:08,370 --> 00:07:12,030
Let's do that on B O question mark T

153
00:07:12,030 --> 00:07:15,900
which is giving us BT and bot.

154
00:07:15,900 --> 00:07:20,900
Whereas if we use the B plus B O plus T

155
00:07:22,620 --> 00:07:26,640
we also see boot and
boot with, with many O's.

156
00:07:26,640 --> 00:07:28,710
Now we can also use a regular expression

157
00:07:28,710 --> 00:07:30,810
that's looking for an exact match.

158
00:07:30,810 --> 00:07:34,650
So I am using grep on BO

159
00:07:34,650 --> 00:07:39,300
and I want to find lines
that have exactly two Os

160
00:07:39,300 --> 00:07:40,650
not more, not less.

161
00:07:40,650 --> 00:07:41,730
How do we do that?

162
00:07:41,730 --> 00:07:44,400
By using slash curly brace,

163
00:07:44,400 --> 00:07:45,540
followed by a two

164
00:07:45,540 --> 00:07:47,160
followed by another slash

165
00:07:47,160 --> 00:07:49,170
followed by another curly brace

166
00:07:49,170 --> 00:07:50,370
followed by a T

167
00:07:50,370 --> 00:07:53,433
and a single quote on users.

168
00:07:54,420 --> 00:07:56,940
So this is giving me an exact match.

169
00:07:56,940 --> 00:08:00,210
Now, if you want to, to
understand it's a little bit

170
00:08:00,210 --> 00:08:01,200
hard to understand,

171
00:08:01,200 --> 00:08:02,850
but we're what it really comes down

172
00:08:02,850 --> 00:08:04,956
to is a two between braces

173
00:08:04,956 --> 00:08:07,680
but these braces need to be escaped.

174
00:08:07,680 --> 00:08:09,630
And that's why there's
this slice before them

175
00:08:09,630 --> 00:08:14,100
and that makes them a
little bit more readable.

176
00:08:14,100 --> 00:08:15,343
But this is how it works.

177
00:08:15,343 --> 00:08:17,670
What happens if you don't escape them?

178
00:08:17,670 --> 00:08:21,360
Well, then you don't get a
result and it's just wrong.

179
00:08:21,360 --> 00:08:23,881
Now there are other regular
expressions as well.

180
00:08:23,881 --> 00:08:28,290
Like for instance, you
are looking for a string

181
00:08:28,290 --> 00:08:32,003
that must be a word grep on slash B

182
00:08:35,700 --> 00:08:39,930
Anna slash B on users

183
00:08:39,930 --> 00:08:40,763
for instance.

184
00:08:40,763 --> 00:08:43,743
And now we can see all
occurrences where Anna is a word

185
00:08:43,743 --> 00:08:47,490
it's not really necessary
to do the B at the beginning

186
00:08:47,490 --> 00:08:52,350
because it works out well in either way.

187
00:08:52,350 --> 00:08:55,770
You also are going to
appreciate the one character.

188
00:08:55,770 --> 00:09:00,770
So if you use grab dot,
dot, dot, dot on users

189
00:09:03,630 --> 00:09:05,790
can you predict what is going to happen?

190
00:09:05,790 --> 00:09:08,340
It is just showing all of the lines.

191
00:09:08,340 --> 00:09:09,173
Why?

192
00:09:09,173 --> 00:09:12,540
Well, because it's looking
for, for characters

193
00:09:12,540 --> 00:09:13,556
any character

194
00:09:13,556 --> 00:09:16,020
that's probably not what you wanted.

195
00:09:16,020 --> 00:09:18,840
If you are looking for
four characters in a word,

196
00:09:18,840 --> 00:09:20,133
for instance,

197
00:09:20,133 --> 00:09:23,250
that's becoming more interesting

198
00:09:23,250 --> 00:09:27,900
then you use slash b behind it.

199
00:09:27,900 --> 00:09:30,480
Then it's considering individual words.

200
00:09:30,480 --> 00:09:31,980
Is that convincing?

201
00:09:31,980 --> 00:09:35,490
Well, maybe not, but if we
put a B in the beginning

202
00:09:35,490 --> 00:09:37,860
then we get something
that really makes sense.

203
00:09:37,860 --> 00:09:41,520
We are looking for four
character words right here

204
00:09:41,520 --> 00:09:44,490
and likewise, you can
use carrot and dollar

205
00:09:44,490 --> 00:09:47,943
as the start and end of the lines.

206
00:09:49,668 --> 00:09:51,750
Last regular expression

207
00:09:51,750 --> 00:09:56,220
I want to show you is another
extended regular expression.

208
00:09:56,220 --> 00:10:01,220
Grep minus uppercase E,
followed by SVM pipe VMX

209
00:10:06,690 --> 00:10:11,610
on slash proc, slash CPU info.

210
00:10:13,710 --> 00:10:16,746
And this is actually a
useful regular expression

211
00:10:16,746 --> 00:10:20,700
because proc CPU info is a
file that exists in the kernel.

212
00:10:20,700 --> 00:10:24,480
And this file contains CPU information.

213
00:10:24,480 --> 00:10:28,659
Let me show you cat on
slash proc slash CPU info.

214
00:10:28,659 --> 00:10:30,240
This is the file.

215
00:10:30,240 --> 00:10:32,430
If you wanna do virtualization

216
00:10:32,430 --> 00:10:37,290
in order to do virtualization,
you need either SVM

217
00:10:37,290 --> 00:10:40,740
or VMX to be in the
result of proc CPU info.

218
00:10:40,740 --> 00:10:44,730
And this is how you can do an
either or regular expression.

219
00:10:44,730 --> 00:10:46,200
Let me just fake it

220
00:10:46,200 --> 00:10:48,660
because here we don't see a result at all.

221
00:10:48,660 --> 00:10:50,220
That's not very convincing.

222
00:10:50,220 --> 00:10:51,450
I'm just going to fake it.

223
00:10:51,450 --> 00:10:54,990
And instead of VMX, I'm
looking for VME and there we go

224
00:10:54,990 --> 00:10:58,930
you can see that that is
actually giving a match.

225
00:10:58,930 --> 00:11:02,190
Now what if we also include PSE

226
00:11:02,190 --> 00:11:04,650
then you can see that
both are giving a match.

227
00:11:04,650 --> 00:11:06,480
This either or regular expression

228
00:11:06,480 --> 00:11:09,840
in fact is a very useful
regular expression.

229
00:11:09,840 --> 00:11:11,250
Are you feeling confused?

230
00:11:11,250 --> 00:11:12,300
I can imagine.

231
00:11:12,300 --> 00:11:15,210
Regular expressions are confusing

232
00:11:15,210 --> 00:11:16,860
and even the best make errors

233
00:11:16,860 --> 00:11:18,780
while using regular expression.

234
00:11:18,780 --> 00:11:21,780
My recommendation, practice
what we have seen right here

235
00:11:21,780 --> 00:11:23,550
and you will get familiar with them

236
00:11:23,550 --> 00:11:25,083
one time, sooner or later.