1
00:00:02,280 --> 00:00:06,000
In this unit we will see how
we can combine commands together,

2
00:00:06,000 --> 00:00:09,560
in order to control
the flow of these commands.

3
00:00:09,560 --> 00:00:10,820
Specifically,

4
00:00:10,820 --> 00:00:13,580
we'll see how we can execute
commands conditionally

5
00:00:13,580 --> 00:00:15,540
with an if or an else,

6
00:00:15,540 --> 00:00:19,400
and also how can create
loops to iterate over files,

7
00:00:19,400 --> 00:00:22,720
or while a condition is true.

8
00:00:22,720 --> 00:00:29,300
The "for" command allows us
to iterate over a set of values.

9
00:00:29,300 --> 00:00:33,260
In order to see the time in all
three European cities, of London,

10
00:00:33,260 --> 00:00:37,240
Paris and Madrid,
I place them in a for loop.

11
00:00:37,240 --> 00:00:40,500
The syntax starts
with the iteration statement:

12
00:00:40,500 --> 00:00:44,300
for i in all these
three timezones; do,

13
00:00:44,300 --> 00:00:46,880
continues with
the commands we specify,

14
00:00:46,880 --> 00:00:52,600
which in our case is to: showtime
the America/Los_Angeles time 11:45

15
00:00:52,600 --> 00:00:56,960
in the $i timezone,
and concludes with a done.

16
00:00:56,960 --> 00:01:02,920
The i variable receives the values
of London, then Paris, and lastly Madrid.

17
00:01:02,920 --> 00:01:07,500
I'm iterating with i through
these values and according to the output:

18
00:01:07,500 --> 00:01:13,140
when the time in L.A. is 11:45,
in London it is 7:45 PM,

19
00:01:13,140 --> 00:01:17,660
whereas in Paris
and Madrid it is 8:45PM.

20
00:01:17,660 --> 00:01:22,240
Apart from simple for loops,
we can also define nested for loops.

21
00:01:22,240 --> 00:01:25,120
For instance, to extend
the previous example,

22
00:01:25,120 --> 00:01:27,380
we can specify
a set of timezones,

23
00:01:27,380 --> 00:01:30,560
and then convert time
between any pair of them.

24
00:01:30,560 --> 00:01:36,840
To do that, I firstly set a variable TZONES
to the timezones I'm interested in:

25
00:01:36,840 --> 00:01:40,200
Los Angeles,
New York and London.

26
00:01:40,200 --> 00:01:44,500
Then I iterate through
the values of TZONES twice;

27
00:01:44,500 --> 00:01:47,520
the first for loop uses
the timezones as input,

28
00:01:47,520 --> 00:01:49,600
and stores them in tzin,

29
00:01:49,600 --> 00:01:54,480
while the second for loop
has them stored as output in tzout.

30
00:01:54,480 --> 00:01:56,340
Using the function showtime,

31
00:01:56,340 --> 00:02:02,580
I now convert the time 11:45
for each tzin to each tzout.

32
00:02:02,580 --> 00:02:10,860
The result we see is a conversion
of 11:45 between all timezones.

33
00:02:10,860 --> 00:02:14,560
The "if" command allows us
to execute various other commands,

34
00:02:14,560 --> 00:02:16,640
based on the result of a condition;

35
00:02:16,640 --> 00:02:22,980
the result of whether another
command succeeded or not.

36
00:02:23,680 --> 00:02:26,600
As a demonstration,
I create two files:

37
00:02:26,600 --> 00:02:29,780
the sourcefile
and the destfile, with a bit of delay,

38
00:02:29,780 --> 00:02:34,460
so that the destfile is clearly
more recent than the sourcefile.

39
00:02:34,460 --> 00:02:36,740
To refresh a file
based on another,

40
00:02:36,740 --> 00:02:39,940
for example to refresh
the destfile based on sourcefile,

41
00:02:39,950 --> 00:02:45,080
I run: if test sourcefile -nt destfile;

42
00:02:45,080 --> 00:02:48,040
using the test command
with the nt option,

43
00:02:48,040 --> 00:02:53,220
to check whether the sourcefile
is newer than the destfile.

44
00:02:53,220 --> 00:02:57,160
In case it is newer,
sourcefile is copied to the destfile,

45
00:02:57,160 --> 00:03:00,520
and I echo that
the destfile has been refreshed.

46
00:03:00,520 --> 00:03:05,540
Apparently, nothing happens because
the sourcefile is older than the destfile,

47
00:03:05,540 --> 00:03:10,520
and, as a result,
the destfile is not refreshed.

48
00:03:10,520 --> 00:03:12,480
To verify the opposite case,

49
00:03:12,480 --> 00:03:16,280
I recreate the sourcefile
to be newer than the destfile

50
00:03:16,280 --> 00:03:18,060
and run the command again,

51
00:03:18,060 --> 00:03:22,080
this time using an alternative
syntax for the test command:

52
00:03:22,080 --> 00:03:25,200
a set of square brackets.

53
00:03:25,200 --> 00:03:28,960
Again I test whether
the sourcefile is newer than the destfile,

54
00:03:28,960 --> 00:03:32,440
and repeat the same
copy and echo commands.

55
00:03:32,440 --> 00:03:36,140
We notice that the destfile
is actually refreshed this time,

56
00:03:36,140 --> 00:03:40,060
because it is older
than the sourcefile.

57
00:03:40,060 --> 00:03:44,300
The if command can also
be combined with an else case.

58
00:03:44,300 --> 00:03:49,820
In a similar example, I test whether
the sourcefile is newer than the destfile,

59
00:03:49,820 --> 00:03:56,580
and after writing the same
copy and echo commands,

60
00:03:58,980 --> 00:04:01,180
I add an else command.

61
00:04:01,180 --> 00:04:06,320
In case the condition is false,
I echo that the destfile is up to date.

62
00:04:06,320 --> 00:04:12,600
As expected, the destfile
is shown to be up to date.

63
00:04:14,540 --> 00:04:21,860
The "while" command allows us
to iterate while a condition holds.

64
00:04:21,860 --> 00:04:26,900
Let's now write a small script that counts
the average number of characters per line,

65
00:04:26,900 --> 00:04:30,940
for all readable
files in the etc directory.

66
00:04:30,940 --> 00:04:36,100
First, I list the directory entries
redirecting the output of ls,

67
00:04:36,100 --> 00:04:38,840
with a pipe
to a while loop.

68
00:04:38,840 --> 00:04:42,440
While takes as an argument
another command, read,

69
00:04:42,440 --> 00:04:47,340
in order to read the input
directory entries line by line.

70
00:04:47,340 --> 00:04:51,660
Each filename is stored
in a variable called name.

71
00:04:51,660 --> 00:04:57,680
Then for each filename, I test with an if
command, if it is a regular readable file,

72
00:04:57,680 --> 00:05:01,780
passing the -f option
to check it's a regular file,

73
00:05:01,780 --> 00:05:07,700
the -a option for and,
and the -r option to check if it's readable.

74
00:05:07,700 --> 00:05:12,260
In case the entry is an actual file
and not a directory, and is readable,

75
00:05:12,260 --> 00:05:18,560
I display the filename without a newline,
by passing the -n option to echo.

76
00:05:18,560 --> 00:05:21,220
Next, I run the expr command,

77
00:05:21,220 --> 00:05:26,060
to calculate the average number of 
characters per line in the particular file.

78
00:05:26,060 --> 00:05:32,560
Specifically, I count the characters
of the file using wc -c and the file as input,

79
00:05:32,560 --> 00:05:38,540
along with the number of its
lines using wc -l, and divide the two.

80
00:05:38,540 --> 00:05:45,340
I complete the if command with fi
and the while loop with done,

81
00:05:45,340 --> 00:05:51,880
and pipe the output to head
to see the first few lines.

82
00:05:52,900 --> 00:05:58,000
The "xargs" command is a magnificent
tool for working with huge datasets.

83
00:05:58,000 --> 00:06:01,060
The command reads
data from its standard input,

84
00:06:01,060 --> 00:06:05,680
and then supplies them as arguments
to a specified command that it executes.

85
00:06:05,680 --> 00:06:08,540
It is needed,
because the operating system,

86
00:06:08,540 --> 00:06:13,580
provides only a fixed amount of space
for providing arguments to a command.

87
00:06:13,580 --> 00:06:18,360
Xargs overcomes this limit
by executing the command repeatedly,

88
00:06:18,360 --> 00:06:24,260
with new batches of
arguments chunked from its input.

89
00:06:24,260 --> 00:06:28,560
Consider as an example, that we
want to find how many lines exist,

90
00:06:28,560 --> 00:06:34,500
in all files under a specific directory,
in my case the current directory.

91
00:06:34,500 --> 00:06:37,280
I run the find command
in the current directory,

92
00:06:37,280 --> 00:06:40,980
specifying with -type f
that I'm interested in files,

93
00:06:40,980 --> 00:06:44,380
and then apply cat
through xargs on all files,

94
00:06:44,380 --> 00:06:47,640
in order to concatenate
the files together.

95
00:06:47,640 --> 00:06:56,020
Finally, I pipe the output of the wc -l 
command, to count the number of lines.

96
00:06:56,020 --> 00:07:00,200
Find doesn't always
work as expected with xargs.

97
00:07:00,200 --> 00:07:03,860
For instance, let's say
I want to find the oldest file,

98
00:07:03,860 --> 00:07:07,820
of the directory
"Program Files" on a Windows machine.

99
00:07:07,820 --> 00:07:14,100
To do that I run find
for "Program Files" and type file,

100
00:07:14,100 --> 00:07:18,920
I apply stat to get
the modification time of these files,

101
00:07:18,920 --> 00:07:25,700
sort through the output numerically,
and display the oldest file using head -1.

102
00:07:25,700 --> 00:07:28,240
But, instead of the oldest file,

103
00:07:28,240 --> 00:07:33,340
we get various errors saying:
cannot find Application or Verifier.

104
00:07:33,340 --> 00:07:36,400
The problem here is
that the output of find,

105
00:07:36,400 --> 00:07:39,920
contains some lines with spaces, 
because on the Windows system,

106
00:07:39,920 --> 00:07:44,780
filenames often have
their words separated by space.

107
00:07:44,780 --> 00:07:49,560
Since xargs uses both a newline
and a space to break up its arguments,

108
00:07:49,560 --> 00:07:53,840
some filenames containing
spaces are incorrectly broken apart,

109
00:07:53,840 --> 00:07:59,280
leading to the creation
and display of non-existing files.

110
00:07:59,280 --> 00:08:03,340
To get around this,
find offers the option -print0,

111
00:08:03,340 --> 00:08:07,020
which separates the output
elements using a null character,

112
00:08:07,020 --> 00:08:08,440
instead of a newline,

113
00:08:08,440 --> 00:08:12,740
based on the fact that
there are no nulls in filenames.

114
00:08:12,740 --> 00:08:17,660
The output of find is then
piped to xargs using a -0 option,

115
00:08:17,660 --> 00:08:21,280
to specify that the input
is separated by null characters,

116
00:08:21,280 --> 00:08:23,960
rather than newlines or spaces.

117
00:08:23,960 --> 00:08:28,640
Again, I output the modification
time and the name of each file,

118
00:08:28,640 --> 00:08:30,780
sorting the results numerically,

119
00:08:30,780 --> 00:08:36,280
and retrieving the oldest file,
which is a LICENCE file of gsview.

120
00:08:36,280 --> 00:08:40,180
By listing the file we verify
that it was created 20 years ago,

121
00:08:40,180 --> 00:08:42,920
in the year 2000.

122
00:08:45,280 --> 00:08:48,840
The "case" command
allows us to run a specific command,

123
00:08:48,840 --> 00:08:52,740
based on pattern matching.

124
00:08:53,540 --> 00:08:56,800
To see this in action
I create some command aliases,

125
00:08:56,800 --> 00:09:00,300
that depend on
the operating system I currently run.

126
00:09:00,300 --> 00:09:02,580
These are useful
to make our commands work,

127
00:09:02,580 --> 00:09:05,900
independently of
the system we use.

128
00:09:05,900 --> 00:09:10,120
The uname command displays
the name of the current operating system,

129
00:09:10,120 --> 00:09:12,540
which in my case is Linux.

130
00:09:12,540 --> 00:09:15,960
With a case statement
I use the output of uname,

131
00:09:15,960 --> 00:09:20,560
and then insert an "in" and some
patterns that specify what happens,

132
00:09:20,560 --> 00:09:23,200
depending on the value of uname.

133
00:09:23,200 --> 00:09:26,780
In case of Linux,
I create a command alias s,

134
00:09:26,780 --> 00:09:30,260
which starts a document
using gnome-open.

135
00:09:30,260 --> 00:09:33,500
Then, I create
a second alias cpt,

136
00:09:33,500 --> 00:09:36,400
that copies the current
directory to the clipboard,

137
00:09:36,400 --> 00:09:43,760
using pwd to print the current
directory, a pipe, and xsel --clipboard.

138
00:09:43,760 --> 00:09:47,940
With a double semicolon
I end the first case.

139
00:09:47,940 --> 00:09:50,580
I then start
another one for Darwin,

140
00:09:50,580 --> 00:09:55,040
which is the output of
uname on an Apple macOS machine.

141
00:09:55,040 --> 00:09:59,240
Similarly the alias s
is set to "open" here,

142
00:09:59,240 --> 00:10:04,580
whereas cpt is set
to "pwd | pbcopy".

143
00:10:04,580 --> 00:10:10,600
On the Cygwin, Unix-like environment
for Windows, the aliases vary accordingly.

144
00:10:10,600 --> 00:10:15,060
Finally, after ending
the case command I run alias,

145
00:10:15,060 --> 00:10:21,060
and verify that the aliases cpt and s
correspond to the Linux case.

146
00:10:21,060 --> 00:10:29,620
So cpt contains "pwd | xsel",
while s is set to "gnome-open".

147
00:10:31,300 --> 00:10:36,760
This concludes our foundations
unit on execution control; stay with us!