Sed: Difference between revisions
m →Introduction: tr |
No edit summary |
||
Line 1: | Line 1: | ||
== Introduction == | == Introduction == | ||
Sed stands for stream editor and it does just that, it edits a stream. There is many uses for it for example in the well known substitution mode. Sometimes [[tr]] does a job better than sed. | Sed stands for stream editor and it does just that, it edits a stream of text. There is many uses for it, for example in the well known substitution mode. Sometimes [[tr]] does a job better than sed. | ||
== sed and substitution == | == sed and substitution == | ||
Line 19: | Line 19: | ||
== sed and deleting == | == sed and deleting == | ||
If I wanted to delete lines 1 through 46 | If I wanted to delete lines 1 through 46: | ||
francisco$ sed -e '1,46d' /etc/passwd | francisco$ sed -e '1,46d' /etc/passwd | ||
Line 29: | Line 29: | ||
== sed and printing certain rows == | == sed and printing certain rows == | ||
To see the first entry in the /etc/passwd file: | |||
francisco$ sed -n 1p /etc/passwd | francisco$ sed -n 1p /etc/passwd | ||
root:*:0:0:Charlie &:/root:/bin/ksh | root:*:0:0:Charlie &:/root:/bin/ksh | ||
And this is how you | And this is how you display rows 2 through 5: | ||
francisco$ sed -n 2,5p /etc/passwd | francisco$ sed -n 2,5p /etc/passwd |
Revision as of 21:47, 13 August 2008
Introduction
Sed stands for stream editor and it does just that, it edits a stream of text. There is many uses for it, for example in the well known substitution mode. Sometimes tr does a job better than sed.
sed and substitution
Using the s allows the results to be substituted like so:
echo one two three four five six seven eight nine | sed -e 's/one/bleep/g' -e 's/five/bleep/g'
would result in:
bleep two three four bleep six seven eight nine
In a directory where files were backed up as "file-save" the renaming from the backup to the original file went like this:
ls *save| while read i ; do NEW=`echo $i | sed -e 's/-save//g'`; mv -f $i $NEW; done
sed and deleting
If I wanted to delete lines 1 through 46:
francisco$ sed -e '1,46d' /etc/passwd _postfix:*:507:507:Postfix Daemon:/var/empty:/sbin/nologin testuser:*:1001:1001:test:/usr/home/testuser:/bin/ksh
You see the last two entries of a 48 line file. This can also be achieved with head.
sed and printing certain rows
To see the first entry in the /etc/passwd file:
francisco$ sed -n 1p /etc/passwd root:*:0:0:Charlie &:/root:/bin/ksh
And this is how you display rows 2 through 5:
francisco$ sed -n 2,5p /etc/passwd daemon:*:1:1:The devil himself:/root:/sbin/nologin operator:*:2:5:System &:/operator:/sbin/nologin bin:*:3:7:Binaries Commands and Source,,,:/:/sbin/nologin smmsp:*:25:25:Sendmail Message Submission Program:/nonexistent:/sbin/nologin
sed delimiters
pretend you see this:
sed -e 's=rat=/raven/=g' file
it looks difficult because you may not know that sed can use any delimiter between a substitute.
In the example given every instance of rat is changed to /raven/ with the slashes. Here is another one
sed -e s,/bin,/newbin,g file
Here the delimiter is a comma (,) and / is part of the substitute text. This is totally ok.