Text files
We now concentrate on a particular class of files, text
files, which are files divided into lines
separated by a Newline character. Such files, which would
normally contain only printable characters, include text program
source files, shell scripts, and in fact any files you would wish
to use a text editor on. However, this is not a requirement, and
what we discuss in this section will also hold true for files
containing other characters. Most of the files we will use as
examples in the rest of this book will be text files.
Suppose we have created a file called story, which
contains English text. Having established that it is a text file
(by means of file or otherwise), we may wish briefly
to examine its contents. We could, of course, invoke an editor such
as vi and use the commands within the editor to move
through the file and look at various parts of it, or we could use a
pager. However, there are easier methods.
Often you will simply want to look at the first few lines of a
file (for instance to verify that it was indeed the file you
expected it to be). In this case, head will print out
the first 5 lines of the file. In a similar vein, tail
will print out the last 5. If you want to see the first (say) 10
lines, then the command would be
head -n 10
where n = 'number'.
Strictly speaking, tail copies its input to
standard output beginning at a designated place, which is
usually a number of lines from the end of the file. There are many
options available to tail to allow you to specify what
is meant by the designated place, and how many lines are output -
refer to the manual page for further details.
As an example of simple use, suppose file myfile
contains 100 lines, as follows:
line 1
line 2
...
line 100
Then we might have
$ head -n 2 myfile
line 1
line 2
$ tail -n 3 myfile
line 98
line 99
line 100
$
Worked example
Find the most recently modified file (excluding 'dot' files) in
the current directory.
Solution: This has clearly somehow got to involve
ls. With option -l we could examine by hand every file
and see which one was last changed. That is not the UNIX way of
doing things - by examining the manual page for ls we
find an option -t ('time') which will sort the files
it prints out so that the most recent is shown first. Option
-1 (1 is digit 'one') forces the output
to be one filename per line. Thus ls -t1 will produce
a list of filenames with the desired one at the top - use
head to isolate it by piping the output of
ls to head thus:
$ ls -t1 | head -n 1
There is only a limited amount of space on a machine, and it may
be that each user has been restricted (by the system administrator)
as to how much filespace they are allowed to use. A quick way to
find out how big files are is to use wc ('wordcount'),
which indicates (i) the number of lines, (ii) the number
of words and (iii) the number of characters
(bytes) in a file. The latter two are only meaningful if the file
is a text file, though. For example,
$ wc myfile
27 124 664 myfile
indicates that file myfile has 27 lines,
124 words and 664 bytes (characters). With
options -c, -w or -l
respectively, only the byte, word or
line count will be printed. Note that wc does
not work on directories.
|