I have a string in a file like this:
"Value1=[random number]"
After this string I want to add another string, specifically:
"Value2=100"
If I try to use:
sed '/Value1=/a Value2=100' myfile.txt
It will fail because I have not included the fact that Value1=[some random number].
How do I add the condition that Value1=random number and Value2 should be added to this string?
sed '/Value1=[0-9]\+/a Value2=100' myfile.txt
The [0-9]\+ will match any string of digits. For example, on my cygwin, GNU sed 4.2.2,
echo Value1=42 | sed '/Value1=[0-9]\+/a Value2=43'
produces
Value1=42
Value2=43
Edit: If the number may or may not be in double-quotes, use:
sed '/Value1="\?[0-9]\+"\?/a Value2=43'
The "\? is an optional double-quote.
Something like this:
$ sed 's,\(Value1=200\),\1 Value2=100,' myfile.txt
Result:
Value1=200 Value2=100
echo '"Value1=400"' | sed 's/"Value1=.*"/&\n"Value2=100"/'
Output:
"Value1=400"
"Value2=100"
Or:
echo '"Value1=400"' | sed 's/"Value1=.*"/& "Value2=100"/'
Output:
"Value1=400" "Value2=100"
Related
I have this data in file.txt:
1234-abca-dgdsf-kds-2;abc dfsfds 2
123-abcdegfs-sdsd;dsfdsf dfd f
12523-cvjbsvndv-dvd-dvdv;dsfdsfpage
I want to replace the string after "-" and up to ";" with just ";", so that I get:
1234;abc dfsfds 2
123;dsfdsf dfd f
12523;dsfdsfpage
I tried with the command:
sed -e "s/-.*;/;" file.txt
But it gives me the following error:
sed command garbled
Why is this happening?
sed replacement commands are defined as (source):
's/REGEXP/REPLACEMENT/[FLAGS]'
(substitute) Match the regular-expression against the content of the pattern space. If found, replace matched string with REPLACEMENT.
However, you are saying:
sed "s/-.*;/;"
That is:
sed "s/REGEXP/REPLACEMENT"
And hence missing a "/" at the end of the expression. Just add it to have:
sed "s/-.*;/;/"
# ^
You are missing a slash at the end of the sed command:
Should be "s/-.*;/;/"
-.* here the * greedy, so this would fail if there are more than one ;
echo "12523-cvjbsvndv-dvd-dvdv;dsfdsfpage;test" | sed -e "s/-.*;/;/"
12523;test
Change to -[^;]*
echo "12523-cvjbsvndv-dvd-dvdv;dsfdsfpage;test" | sed -e "s/-[^;]*;/;/"
12523;dsfdsfpage;test
This should work :
sed 's/-.*;/;/g' file > newFile
Sed replaces with carriage returns.
My file:
<ABC>123</ABC><ABC>456</ABC>
I want to output like this:
<ABC>123</ABC>
<ABC>456</ABC>
I am using:
sed "s/<ABC>///n/g"
But there isn't any changes.
Try as below
sed -e 's/></>\n</g' file.txt > output.txt
Kindly help. I want to make .0 to be 0.0 within a UNIX file yyyyy.csv :
603905209;47.824;USD
603905477;57.199;USD
603938657;3.2281;USD
603949388;.00191;USD
603937274;.00563;USD
603911160;.00287;USD
I want the result to be
603905209;47.824;USD
603905477;57.199;USD
603938657;3.2281;USD
603949388;0.00191;USD
603937274;0.00563;USD
603911160;0.00287;USD
but I got this result:
603905209;0.4.7824;USD
603905477;0.5.7199;USD
603938657;0.3.2281;USD
603949388;0.00191;USD
603937274;0.00563;USD
603911160;0.00287;USD
Below is my command:
sed 's/;.0/;0.0/g' yyyyy.csv | sed 's/;.2/;0.2/g' | sed 's/;.1/;0.1/g' | sed 's/;.3/;0.3/g' | sed 's/;.4/;0.4/g' | sed 's/;.5/;0.5/g'| sed 's/;.6/;0.6/g' | sed 's/;.7/;0.7/g' | sed 's/;.8/;0.8/g' | sed 's/;.9/;0.9/g' > xxxxx.csv
You need to escape all the dots present in your regex or otherwise it would match any character. That is, . is a special meta character in regex which matches any character. To match a literal dot, you need to escape the ..
sed 's/;\.0/;0.0/g' yyyyy.csv
And this would be enough.
$ sed 's/;\.\([0-9]\)/;0.\1/g' file
603905477;57.199;USD
603938657;3.2281;USD
603949388;0.00191;USD
603937274;0.00563;USD
603911160;0.00287;USD
In basic sed, \(...\) called capturing group, which is used to capture the characters matched by the pattern present inside that group. So the pattern present inside the group is [0-9] which matches a digit from 0-9. We could refer the captured characters through back-referencing ie, \1. \1 at the replacement part refers to the characters which are present inside the group index 1.
Change your command to:
sed 's/;\./;0./g' File
i.e, just substitute ;. with ;0..
Just print your file using a tool that understands numbers:
$ awk -F';' '{printf "%d;%f;%s\n", $1,$2,$3}' file
603905209;47.824000;USD
603905477;57.199000;USD
603938657;3.228100;USD
603949388;0.001910;USD
603937274;0.005630;USD
603911160;0.002870;USD
awk -F';' '{printf "%d;%.5f;%s\n", $1,$2,$3}' file
603905209;47.82400;USD
603905477;57.19900;USD
603938657;3.22810;USD
603949388;0.00191;USD
603937274;0.00563;USD
603911160;0.00287;USD
$ awk -F';' '{printf "%d;%.2f;%s\n", $1,$2,$3}' file
603905209;47.82;USD
603905477;57.20;USD
603938657;3.23;USD
603949388;0.00;USD
603937274;0.01;USD
603911160;0.00;USD
Whatever precision you like...
I am new to sed . I want to replace a substring
for example:
var1=server1:game1,server2:game2,sever3:game1
output should be
server1 server2 server3 (with just spaces)
I have tried this.
echo $var1 | sed 's/,/ /g' | sed 's/:* / /g'
This is not working. Please suggest a solution.
You can try this sed,
echo $var1 | sed 's/:[^,]\+,\?/ /g'
Explanation:
:[^,]\+, - It will match the string from : to ,
\? - Previous may occur or may not ( Since end of line don't have , )
echo $var1 | sed s/:game[0-9],*/\ /
Assuming your sub string has game followed by a number([0-9]*)
An awk variation using same regex as sed
awk '{gsub(/:[^,]+,?/," ")}1' <<< "$var1"
PS Its always good custom to "quote" variables
Just for info, you are really only matching, not replacing, so grep can be your friend (with -P):
grep -oP '[^:,=]+(?=:)'
That matches a number of characters that aren't :,= followed by a : using lookahead.
This will put the servers on different lines, which may be what you want anyway. You can put them on one line by adding tr:
grep -oP '[^:,=]+(?=:)' | tr '\n' ' '
I'm having a problem trying to filter the following string in unix
<option value="20121209/YvegRascYTGxmWLUIrqW/por121209130030.jpg">2012-12-09 13:00h</option>
into:
20121209/YvegRascYTGxmWLUIrqW/por121209130030.jpg and 2012-12-09 13:00.
I can find the beginning of the substrings I want, but not the ending.
file=tmpfile
read -r firstline<$file
firstArg=$(echo $firstline | sed 's/^.*value="//' | sed 's/">*$//')
echo $firstArg
secondArg=$(echo $firstline | sed 's/^.*">//' | sed 's/h<*$//')
echo $secondArg
The output is the following:
20121209/YvegRascYTGxmWLUIrqW/por121209130030.jpg">2012-12-09 13:00h</option>
2012-12-09 13:00h</option>
But what i really want is
20121209/YvegRascYTGxmWLUIrqW/por121209130030.jpg
2012-12-09 13:00
Minor fix, you are missing . in two places (before the * of the 2nd sed on each line) :
file=tmpfile
read -r firstline<$file
firstArg=$(echo $firstline | sed 's/^.*value="//' | sed 's/">.*$//')
echo $firstArg
secondArg=$(echo $firstline | sed 's/^.*">//' | sed 's/h<.*$//')
echo $secondArg
Input:
<option value="20121209/YvegRascYTGxmWLUIrqW/por121209130030.jpg">2012-12-09 13:00h</option>
Output:
20121209/YvegRascYTGxmWLUIrqW/por121209130030.jpg
2012-12-09 13:00
Explanation:
* matches the preceding character 0 or more times, so you were matching zero or more >'s and <'s previously.
The . matches any character
So >.* matches zero or more characters after >.
Improvement:
Additionally, the two sed lines can be better written as:
firstArg=$(sed 's/^.*value="//;s/">.*$//' <<< "$firstline")
secondArg=$(sed 's/^.*">//;s/h<.*$//' <<< "$firstline")
Separate multiple replace patterns by ; within one sed call
<<< notation is called a herestring, you can save echos and pipes here by using it instead
Always wrap variables with double quotes
Try doing this :
sed -r 's#.*([0-9]{8}/[^"]+).*>([0-9]{4}-[0-9]{2}-[0-9]{2} [0-9]{2}:[0-9]{2}[[:alnum:]]*)<.*#\1\n\2#g'
EXAMPLE
$ cat file.txt
<option value="20121209/YvegRascYTGxmWLUIrqW/por121209130030.jpg">2012-12-09 13:00h</option>
$ sed -r 's#.*([0-9]{8}/[^"]+).*>([0-9]{4}-[0-9]{2}-[0-9]{2} [0-9]{2}:[0-9]{2}[[:alnum:]]*)<.*#\1\n\2#g' file.txt
20121209/YvegRascYTGxmWLUIrqW/por121209130030.jpg
2012-12-09 13:00h