I am curious how can I do the following thing with sed: I need to wrap certain two words in single quotes in each line. For example, the following text:
I did it like this:
cat file |sed "s/[^ ][^ ]*/'&'/g"
result:
'the-first-word' '-firstPrameter' 'value' '-secondParameter' 'value' '-thirdParameter' 'value' '-fourthParameter' 'value'
But it is wrapping every word.
It should be like this, skip the first word and wrap only parameter and value:
the-first-word '-firstPrameter value' '-secondParameter value' '-thirdParameter value' '-fourthParameter value'
sed "s/ \([^ ]* [^ ]*\)/ '\\1'/g" file
$cat file.txt
the-first-word -firstPrameter value -secondParameter value -thirdParameter value -fourthParameter value
$sed "s/ \([^ ]* [^ ]*\)/ '\\1'/g" file.txt
the-first-word '-firstPrameter value' '-secondParameter value' '-thirdParameter value' '-fourthParameter value'
In your example, the first word is at the start of the line. Thus, if we quote all words that have a space in front of them, then we will quote all but the first:
sed "s/ \([^ ]\+\)/ '\\1'/g" file
the-first-word '-firstPrameter' 'value' '-secondParameter' 'value' '-thirdParameter' 'value' '-fourthParameter' 'value'
If your sed supports the -r flag (GNU), then we can remove some of those backslashes:
sed -r "s/ ([^ ]+)/ '\\1'/g" file
On Mac OSX, replace -r with -E.
Related
I have an text file, with line inside...
line: <version="AAA" email="...ANY..." file="BBB">
new desired line in text file to be: <version="AAA" email="NEW_TEXT" file="BBB">
I want to replace the ...ANY... expression with variable (replace entire line)
I have this script text-file script in #!/bin/bash, but I have problem when expanding double quotes in variables.
LINE_NUMBER="$(grep -nr 'email' *.txt | awk '{print $1}' | sed 's/[^0-9]*//g')"
VAR1="$(grep 'email' *.txt | cut -d '"' -f1-3)"
VAR2="$(grep 'email' *.txt | cut -d '"' -f5-)"
VAR3='NEW_TEXT'
NEW_LINE=$VAR1'"'$VAR3'"'$VAR2
new desired line in text file to be... <version="AAA" email="NEW_TEXT" file="BBB">
awk -i inplace 'NR=='"$LINE_NUMBER"'{sub(".*",'"'$NEW_LINE'"')},1' *.txt
but I get this new line:
<version="" email="NEW_TEXT" file="">
what do I do wrong?
How can I prevent expand duouble quotes inside variable?
please better write me an working example, I had tried other topics, forums, posts....but I have no luck.
You cas use sed :
VAR3='NEW_TEXT'
sed -i "s/email=\"[^\"]*\"/email=\"$VAR3\"/" myfile.xml
Suggesting:
var3="text space % special < chars"
Note var3 may not contain & which is special replacement meaning in sed
sed -E 's|email="[^"]*"|email="'"${var3}"'"|' input.1.txt
Explanation
[^"]* : Match longest string not having " till next ".
I'm lost trying to do following substitution with sed:
# edit: to capture the full complexity of my problem,
I added the fact that filenames are contained in variables afterwards.
Solutions might therefore directly use the filenames.
given a variable I='insert.txt':
'insert.txt':
Text I wanna skip.
This is text to insert containing
spaces and new lines
given a variable M='toModify.txt':
'toModify.txt':
Insert the new text: here.
I would like to replace the 'here' from $M with the content
of $I:
Insert the new text: This is text to insert containing
spaces and new lines.
I tried:
sed -e "s/here/$(tail -n2 $I | sed -e 's/ /\\ /g' | tr '\n' '\\n')/" $M
with error:
sed unterminated `s' command
The problem is that I don't get the spaces and new lines without terminating the s command.
Any solution?
You can't replace one character with two with tr. Escaping the individual spaces is pointless anyway. The reason for the immediate error is that you end up escaping the final slash, too:
linux$ tail -n2 "$I" | sed -e 's/ /\\ /g' | tr '\n' '\\n'
This\ is\ text\ to\ insert\ containing\spaces\ and\ new\ lines\/
Escaping the spaces is pointless anyway. I guess you want something like this:
linux$ sed '1,2d;$!s/$/\\/' "$I"
This is text to insert containing\
spaces and new lines
We delete lines 1 and two; then add a backslash before every newline except the last.
linux$ sed -e "s/here/$(sed '1,2d;$!s/$/\\/' "$I")/" "$M"
Insert the new text: This is text to insert containing
spaces and new lines.
This is one detail of sed which isn't entirely portable. But the above works for me on Linux and MacOS. (Note you might need to set +H to disable csh-style history expansion aka -bash: !s/$/\\/': event not found errors).
You may use this awk:
awk 'BEGIN{prs=RS; RS=""; getline s < "insert.txt"; RS=prs}
{gsub(/here/, s)} 1' toModify.txt
Insert the new text: This is text to insert containing
spaces and new lines.
Using Perl one-liner
> cat insert.txt
This is text to insert containing
spaces and new lines
> cat toModify.txt
Insert the new text: here
> export I=insert.txt
> export M=toModify.txt
> perl -ne 'BEGIN{$x=qx(cat $ENV{M});$x=~s/here/qx(cat $ENV{I})/e; print $x;exit }'
Insert the new text: This is text to insert containing
spaces and new lines
>
this question originated from string pattaren-matching using awk , basically we are splitting a line of text in multiple groups based on a regex pattern, and then printing two groups only. Now the question is can we right align a group while printing through sed?
below is an example
$cat input.txt
it is line one
it is longggggggg one
itttttttttt is another one
now
$sed -e 's/\(.*\) \(.*\) \(.*\) \(.*\)/\1 \3/g' input.txt
it splits and prints group 1 and 3, but the output is
it line
it longggggggg
itttttttttt another
my question is can we do it through sed so that the output comes as
it line
it longggggggg
itttttttttt another
I did it with awk but I feel it can be done through sed, but I am not able to get how I am going to get the length of the second group and then pad correct number of spaces in between the groups, I am open to any suggestions to try out.
This might work for you (GNU sed):
sed -r 's/^(.*) .* (.*) .*$/\1 \2/;:a;s/^.{1,40}$/ &/;ta;s/^( *)(\S*)/\2\1/' file
or:
sed -r 's/^(.*) .* (.*) .*$/printf "%-20s%20s" \1 \2/e' file
You can use looping in sed to achieve what you want:
#!/bin/bash
echo 'aa bb cc dd
11 22 33333333 44
ONE TWO THREEEEEEEEE FOUR' | \
sed -e 's/\(.*\) \(.*\) \(.*\) \(.*\)/\1 \3/g' \
-e '/\([^ ]*\) \([^ ]*\)/ { :x ; s/^\(.\{1,19\}\) \(.\{1,19\}\)$/\1 \2/g ; tx }'
The two 19's control the width of your columns. The :x is a label which is looped to by tx whenever the preceding substitution succeeded. (You could add a p; before tx to "debug" it.
It most easy to use awk in this case...
You could too use a bash loop to calculate the number of space and run this command on the line covered :
while read; do
# ... calculate $SPACE ...
echo $REPLY|sed "s/\([^\ ]*\)\ *[^\ ]*\ *\([^\ ]*\)/\1$SPACES\2/g"
done < file
But I prefer use awk for do all that (or other advanced shell languages such as Perl, Python, PHP shell mode, ...)
TemplateSpace=" "
TemplateSize=${#TemplateSpace}
sed "
# split your group (based on word here but depend on your real need)
s/^ *\(\w\) \(\w\) \(\w\) \(\w\).*$/\1 \3/
# align
s/$/${TemplateSpace}/
s/^\(.\{${TemplateSize}\}\).*$/\1/
s/\(\w\) \(\w\)\( *\)/\1 \3\2/
"
or more simple for avoiding TemplateSize (and there are no dot in content)
TemplateSpace="............................................................."
and replace
s/^\(.\{${TemplateSize}\}.*$/\1/
by
s/^\(${TemplateSpace}\).*$/\1/
s/\./ /g
Del columns 2 and 4. Right justify resulting col 2 at line length of 23 chars.
sed -e '
s/[^ ]\+/ /4;
s/[^ ]\+//2;
s/^\(.\{23\}\).*$/\1/;
s/\(^[^ ]\+[ ]\+\)\([^ ]\+\)\([ ]\+\)/\1\3\2/;
'
or gnu sed with extended regex:
sed -r '
s/\W+\w+\W+(\w+)\W+\w+$/\1 /;
s/^(.{23}).*/\1/;
s/(+\W)(\w+)(\W+)$/\1\3\2/
'
This question is old, but I like to see it as a puzzle.
While I love the loop solution for its brevity, here is one without a loop or shell help.
sed -E "s/ \w+ (\w+) \w+$/ \1/;h;s/./ /g;s/$/# /;s/( *)#\1//;x;H;x;s/\n//;s/^( *)(\w+)/\2\1/"
or without extended regex
sed "s/ .* \(.*\) .*$/ \1/;h;s/./ /g;s/$/# /;s/\( *\)#\1//;x;H;x;s/\n//;s/^\( *\)\([^ ]*\)/\2\1/"
I want to get the string between <sometag param=' and '>
I tried to use the method from Get any string between 2 string and assign a variable in bash to get the "x":
echo "<sometag param='x'><irrelevant stuff='nonsense'>" | tr "'" _ | sed -n 's/.*<sometag param=_\(.*\)_>.*/\1/p'
The problem (apart from low efficiency because I just cannot manage to escape the apostrophe correctly for sed) is that sed matches the maximum, i.e. the output is:
x_><irrelevant stuff=_nonsense
but the correct output would be the minimum-match, in this example just "x"
Thanks for your help
You are probably looking for something like this:
sed -n "s/.*<sometag param='\([^']*\)'>.*/\1/p"
Test:
echo "<sometag param='x'><irrelevant stuff='nonsense'>" | sed -n "s/.*<sometag param='\([^']*\)'>.*/\1/p"
Results:
x
Explanation:
Instead of a greedy capture, use a non-greedy capture like: [^']* which means match anything except ' any number of times. To make the pattern stick, this is followed by: '>.
You can also use double quotes so that you don't need to escape the single quotes. If you wanted to escape the single quotes, you'd do this:
-
... | sed -n 's/.*<sometag param='\''\([^'\'']*\)'\''>.*/\1/p'
Notice how that the single quotes aren't really escaped. The sed expression is stopped, an escaped single quote is inserted and the sed expression is re-opened. Think of it like a four character escape sequence.
Personally, I'd use GNU grep. It would make for a slightly shorter solution. Run like:
... | grep -oP "(?<=<sometag param=').*?(?='>)"
Test:
echo "<sometag param='x'><irrelevant stuff='nonsense'>" | grep -oP "(?<=<sometag param=').*?(?='>)"
Results:
x
You don't have to assemble regexes in those cases, you can just use ' as the field separator
in="<sometag param='x'><irrelevant stuff='nonsense'>"
IFS="'" read x whatiwant y <<< "$in" # bash
echo "$whatiwant"
awk -F\' '{print $2}' <<< "$in" # awk
Is there a way to send the back references of the SED s/// command to pipes? In order to get an entry from the text, change it, and then write it back. I found that a substitution inside of SED works:
$ echo 'Test ....' | sed 's/Test/'$( echo "<\0>" )'/'
<Test> ....
But the first pipe does not:
$ echo 'Test ....' | sed 's/Test/'$( echo "<\0>" | tr 's' 'x' )'/'
<Test> ....
What is the reason? Additionally, I can't understand why this works at all. I thought that $() substitution should be processed before sed (all the more as I broke the quotes).
And how can I insert one s/// command into another using sed? I use bash.
The tr command is operating on the text "<\0>", not on "<Test>". The back reference isn't expanded in sed until after the pipeline completes. Your second example is equivalent to
foo=$( echo "<\0>" | tr 's' 'x' )
echo 'Test ....' | sed 's/Test/'$foo'/'
It's a little easier to see here that tr has no way of seeing "Test" in its input.
You can achieve the effect you're after with GNU sed and the e flag:
echo 'Test ....' | sed 's/Test.*/echo "<\0>" | tr s x/e'
Output:
<Text ....>