I have strings called:
abc.out
def.out
How do I delete the substring
.out
In these strings?
What command should I use? (Bourne Shell)
Add a comment
|
6 Answers
Multiple ways, a selection:
str=abc.out
Shell:
echo ${str%.*}
Grep:
echo $str | grep -o '^[^\.]*'
Sed:
echo $str | sed -E 's/(.*?)\..*/\1/'
Awk:
echo $str | awk -F. '{print $1}'
-F. means split the string by . and $1 means the first column.
Cut:
echo $str | cut -d. -f1
All output:
abc
Comments
I found this worked best because the pattern you want to use can be in a variable:
DATA="abc.out"
pattern=".out"
DATA=${DATA/$pattern/}
echo "DATA=${DATA}"
The result is:
abc
1 Comment
Sasho
Works for prefixes as well :) Thank you!
If these strings are stored in a file (let's call it input_file):
# input_file:
abc.out abc.out abc.out
def.out def.out
def.out
You can do:
sed -i 's/\.out//g' input_file
And this will remove any occurrence of the substring .out from that file.
Explanation:
sed: invoke the sed tool to edit streams of text-i: use the "in-place" option - this modifies the input file you provide it instead of writing output tostdout's/\.out//g': Use regular expression to delete.out. thegat the end means delete all occurrences.input_file: specify the input file
If these strings are stored in variables:
var1="abc.out"
You can use parameter subsitution:
var1=${var1%.out}
echo "$var1"
abc
Explanation:
- From the above link: "${var%Pattern} Remove from $var the shortest part of $Pattern that matches the back end of $var."
- Note that the "pattern" mentioned here is called globbing, which is different from regular expression in important ways.
2 Comments
William Pursell
You need an end of line anchor on your sed versions, and the
g flag is probably undesirable.
sampson-chen
@WilliamPursell I guess that depends on exact input of his files: you are right, I should make my assumptions more clear.
$ foo=abc.def.out
$ echo ${foo%.out}
abc.def
In general, to delete the shortest suffix, use:
$ echo ${foo%.*}
abc.def
To delete the longest suffix, use
$ echo ${foo%%.*}
abc
3 Comments
not2savvy
Removes only suffixes, not substrings as was asked for.
William Pursell
@not2savvy The OP used the word "substring", but is clearly asking only to delete suffixes. Shall we take all questions literally? I estimate that at least 85% of all questions asked are stated incorrectly. If answered literally, the answers would all be useless.
not2savvy
I think he’s not „clearly asking only to delete suffixes“, bzt he just gave that as one example of what he’s asking for. Perhaps I’m wrong. However, an answer that gives a more general approach is obviously useful for more people. Just saying!
If these are filenames, you can use basename.
$ basename a.out .out
will give you:
a
1 Comment
j--
basename -s .out a.out works too if you like writing your options.An example in zsh:
% myVar="abc.out def.out ghi.out"
% myVar=${myVar:s/.out/} # replace first match of ".out" with ""
% echo myVar
abc def.out ghi.out
% myVar=${myVar:gs/.out/} # replace all matches of ".out" with ""
% echo myVar
abc def ghi
