How would I go about deleting all the lines before the last occurrence of a string. Like if I had a file that looked like
Icecream is good
And
Chocolate is good
And
They have lots of sugar
If I want all lines after and including the last occurrence of "And" what's the cleanest way to do this? Specifically, I want
And
They have lots of sugar
I was doing sed -n -E -e '/And/,$p' file but I see this gives me the first occurrence.
This might work for you (GNU sed):
sed -n '/And/h;//!H;$!d;x;//p' file
Replace anything in the hold space by the line containing And. Append all other lines to the hold space. At the end of the file, swap the pattern space for the hold space and print out the result as long it matches the required string And.
I know that you asked for sed and that Potong provided a good sed solution. But, for comparison, here is an awk solution:
$ awk 's{s=s"\n"$0;} /And/{s=$0;} END{print s;}' file
And
They have lots of sugar
How it works:
s{s=s"\n"$0;}
If the variable s is not empty, then add to it the current line, $0.
/And/{s=$0;}
If the current line contains And, then set s to the current line, $0.
END{print s;}
After we have reached the end of the file, print s.
$ tac file | awk '!f; /And/{f=1}' | tac
And
They have lots of sugar
$ awk 'NR==FNR{if(/And/)nr=NR;next} FNR>=nr' file file
And
They have lots of sugar
Related
I need to align decimal numbers with the "," symbol using only the sed command. The "," should go in the 5th position. For example:
183,7
2346,7
7,999
Should turn into:
183,7
2346,7
7,999
The maximum amount of numbers before the comma is 4. I have tried using this to remove spaces:
sed 's/ //g' input.txt > nospaces.txt
And then I thought about adding spaces depending on the number of digits before the comma, but I don't know how to do this using only sed.
Any help would be appreciated.
Assuming that there is only one number on each line; that there are at most four digits before the ,, and that there is always a ,:
sed 's/[^0-9,]*\([0-9]\+,[0-9]*\).*/ \1/;s/.*\(.....,.*\)/\1/;'
The first s gets rid of everything other than the (first) number on the line, and puts four spaces before it. The second one deletes everything before the fifth character prior to the ,, leaving just enough spaces to right justify the number.
The second s command might mangle input lines which didn't match the first s command. If it is possible that the input contains such lines, you can add a conditional branch to avoid executing the second substitution if the first one failed. With Gnu sed, this is trivial:
sed 's/[^0-9,]*\([0-9]\+,[0-9]*\).*/ \1/;T;s/.*\(.....,.*\)/\1/;'
T jumps to the end of the commands if the previous s failed. Posix standard sed only has a conditional branch on success, so you need to use this circuitous construction:
sed 's/[^0-9,]*\([0-9]\+,[0-9]*\).*/ \1/;ta;b;:a;s/.*\(.....,.*\)/\1/;'
where ta (conditional branch to a on success) is used to skip over a b (unconditional branch to end). :a is the label referred to by the t command.
if you change your mind, here is an awk solution
$ awk -F, 'NF{printf "%5d,%-d\n", $1,$2} !NF' file
183,7
2346,7
7,999
set the delimiter to comma and handle both parts as separate fields
Try with this:
gawk -F, '{ if($0=="") print ; else printf "%5d,%-d\n", $1, $2 }' input.txt
If you are using GNU sed, you could do as below
sed -r 's/([0-9]+),([0-9]+)/printf "%5s,%d" \1 \2/e' input.txt
This may be a better task for awk than sed, but the goal is to parse a single, long string (it happens to be an XML doc) and replace text within a pattern range with another character.
I want to preserve the number of characters being replaced and simply mask them as asterisks. I've put something together in a python script to parse the XML tree but have a feeling a native program is going to be much faster.
Assuming the string: "<mask>123</mask><keep>123</keep>"
...I'd like the output: "<mask>***</mask><keep>123</keep>"
My first attempt with sed without using ranges got me this:
$ echo "<mask>123</mask><keep>123</keep>" | sed "s/[0-9]/*/g"
<mask>***</mask><keep>***</keep>
I learned that sed can operate within ranges, but my understanding is that the behavior can only be toggled from line-to-line, not over the course of processing a single line.
Experimenting with pattern ranges got me the following (consistent with my understanding) and thus didn't work either:
$ echo "<mask>123</mask><keep>123</keep>" | sed "/<mask>/,/<\/mask>/ s/[0-9]/*/g"
<mask>***</mask><keep>***</keep>
EDIT: In fact, even if there were line breaks in the input, I must not be understanding the pattern range behavior correctly (or my example is poorly constructed)
$ echo "<mask>123</mask>\n<keep>123</keep>" | sed "/<mask>/,/<\/mask>/ s/[0-9]/*/g"
<mask>***</mask>
<keep>***</keep>
Any tips would be greatly appreciated.
Never use range expressions as they make simple tasks very slightly briefer but then need a complete rewrite or duplicate conditions when your requirements become marginally more interesting, always use a flag variable instead if a range is necessary. What that means, of course, is that you can't use sed for problems like this since it doesn't support variables.
Anyway, here's a trivial GNU awk (for multi-char RS and RT) solution that doesn't directly use ranges at all:
$ cat file
Assuming the string: "<mask>123</mask><keep>123</keep>" ...I'd like the
$ awk -v RS='</mask>' -v ORS= '{print gensub(/(.*<mask>).*/,"\\1***",1) RT}' file
Assuming the string: "<mask>***</mask><keep>123</keep>" ...I'd like the
or if you need the number of *s to match the number of characters they're replacing:
$ cat file
Assuming first string: "<mask>123</mask><keep>123</keep>" ...I'd like the
Assuming second string: "<mask>1234567</mask><keep>123</keep>" ...I'd like the
$ awk -v RS='</mask>' 'match($0,/(.*<mask>)(.*)/,a){ $0=a[1] gensub(/./,"*","g",a[2]) } {ORS=RT} 1' file
Assuming first string: "<mask>***</mask><keep>123</keep>" ...I'd like the
Assuming second string: "<mask>*******</mask><keep>123</keep>" ...I'd like the
why you got this output is completely correct. It is a trick of sed's range address of two regex.
What you gave sed is /regex1/, /regex2/, sed will first try to find the line matches address1, which is /regex1/, the first line matched, fine. Then your address2 is a regex too, so:
and if addr2 is a regexp, it will not be tested against the line
that addr1 matched.
This sentence is from sed's man page.
That is, sed starts checking your /regex2/ from line 2. of course, no line matches the /<\/mask>/, so sed just did the substitution on whole file.
Check this example:
kent$ cat f
<mask>234</mask>
123
123
123
<mask>234</mask>
123
123
<keep>234</keep>
kent$ sed "/<mask>/,/<\/mask>/ s/[0-9]/*/g" f
<mask>***</mask>
***
***
***
<mask>***</mask>
123
123
<keep>234</keep>
Finally just a suggestion, don't process xml with regex (sed/awk/grep...). Of course, you may just use the "xml" as an example.
I'm working in bash trying to use sed substitution on a file and show both the line number where the substitution occurred and the final version of the line. For a file with lines that contain foo, trying with
sed -n 's/foo/bar/gp' filename
will show me the lines where substitution occurred, but I can't figure out how to include the line number. If I try to use = as a flag to print the current line number like
sed -n 's/foo/bar/gp=' filename
I get
sed: -e expression #1, char 14: unknown option to `s'
I can accomplish the goal with awk like
awk '{if (sub("foo","bar",$0)){print NR $0}}' filename
but I'm curious if there's a way to do this with one line of sed. If possible I'd love to use a single sed statement without a pipe.
I can't think of a way to do it without listing the search pattern twice and using command grouping.
sed -n "/foo/{s/foo/bar/g;=;p;}" filename
EDIT: mklement0 helped me out there by mentioning that if the pattern space is empty, the default pattern space is the last one used, as mentioned in the manual. So you could get away with it like this:
sed -n "/foo/{s//bar/g;=;p;}" filename
Before that, I figured out a way not to repeat the pattern space, but it uses branches and labels. "In most cases," the docs specify, "use of these commands indicates that you are probably better off programming in something like awk or Perl. But occasionally one is committed to sticking with sed, and these commands can enable one to write quite convoluted scripts." [source]
sed -n "s/foo/bar/g;tp;b;:p;=;p" filename
This does the following:
s/foo/bar/g does your substitution.
tp will jump to :p iff a substitution happened.
b (branch with no label) will process the next line.
:p defines label p, which is the target for the tp command above.
= and p will print the line number and then the line.
End of script, so go back and process the next line.
See? Much less readable...and maybe a distant cousin of :(){ :|:& };:. :)
It cannot be done in any reasonable way with sed, here's how to really do it clearly and simply in awk:
awk 'sub(/foo/,"bar"){print NR, $0}' filename
sed is an excellent tool for simple substitutions on a single line, for anything else use awk.
I've got a the following variable LINES with the format date;album;song;duration;singer;author;genre.
August 2013;MDNA;Falling Free;00:31:40;Madonna;Madonna;Pop
August 2013;MDNA;I don't give a;00:45:40;Madonna;Madonna;Pop
August 2013;MDNA;I'm a sinner;01:00:29;Madonna;Madonna;Pop
August 2013;MDNA;Give Me All Your Luvin';01:15:02;Madonna;Madonna;Pop
I want to output author-song, so I made this script:
echo $LINES | sed s_"^[^;]*;[^;]*;\([^;]*\);[^;]*;[^;]*;\([^;]*\)"_"\2-\1"_g
The desired output is:
Madonna-Falling Free
Madonna-I don't give a
Madonna-I'm a sinner
Madonna-Give Me All Your Luvin'
However, I am getting this:
Madonna-Falling Free;Madonna;Pop August 2013;MDNA;I don't give a;00:45:40;Madonna;Madonna;Pop August 2013;MDNA;I'm a sinner;01:00:29;Madonna;Madonna;Pop August 2013;MDNA;Give Me All Your Luvin';01:15:02;Madonna;Madonna;Pop
Why?
EDIT: I need to use sed.
When I run your sed script on your input, I get this output:
Madonna-Falling Free;Pop
Madonna-I don't give a;Pop
Madonna-I'm a sinner;Pop
Madonna-Give Me All Your Luvin';Pop
which is fine except for the extra ;Pop - you just need to add .*$ to the end of your regex so that the entire line is replaced.
Based on your reported output, I'm guessing your input file is using a different newline convention from what sed expects.
In any case, this is a pretty silly thing to use sed for. Much better with awk, for instance:
awk 'BEGIN {FS=";";OFS="-"} {print $5,$3}'
or, slightly more tersely,
awk -F\; -vOFS=- '{print $5,$3}'
If you want sed to see more than one line of input, you must quote the variable to echo:
echo "$LINES" | sed ...
Note that I'm not even going to try to evaluate the correctness of your sed script; using sed here is a travesty, given that awk is so much better suited to the task.
It looks like sed is viewing your entire sample text as a single line. So it is performing the operation requested and then leaving the rest unchanged.
I would look into the newline issue first. How are you populating $LINES?
You should also add to the pattern that seventh field in your input (genre), so that the expression actually does consume all of the text that you want it to. And perhaps anchor the end of the pattern on $ or \b (word boundary) or \s (a spacey character) or \n (newline).
If your format is absolutely permanent, just try below:
echo $line | sed 's#.*;.*;\(.*\);.*;.*;\(.*\);.*#\2-\1#'
I'm running sed as a part of a shell script to clean up bind logs for insertion into a database.
One of the sed commands is the following:
sed -i 's/-/:/g' $DPath/named.query.log
This turns out to be problematic as it disrupts any resource requests that also include a dash (I'm using : as a delimiter for an awk statement further down).
My question is how do I limit the sed command above to only the first ten characters of the line? I haven't seen a specific switch that does this, and I'm nowhere near good enough with RegEx to even start on developing one that works. I can't just use regex to match the preceding numbers because it's possible that the pattern could be part of a resource request. Heck, I can't even use pattern matching for ####-##-## because, again, it could be part of the resource.
Any ideas are much appreciated.
It's [almost always] simpler with awk:
awk '{target=substr($0,1,10); gsub(/-/,":",target); print target substr($0,11)}' file
I think the shortest solution, and perhaps the simplest, is provided by sed itself, rather than awk[ward]:
sed "h;s/-/:/g;G;s/\(..........\).*\n........../\1/"
Explanation:
(h) copy everything to the hold space
(s) do the substitution (to the entire pattern space)
(G) append the hold space, with a \n separator
(s) delete the characters up to the tenth after the \n, but keep the first ten.
Some test code:
echo "--------------------------------" > foo
sed -i "h;s/-/:/g;G;s/\(..........\).*\n........../\1/" foo
cat foo
::::::::::----------------------
I'm not sure how make sed do it per se, however, I do know that you can feed sed the first 10 characters then paste the rest back in, like so:
paste -d"\0" <(cut -c1-10 $DPath/named.query.log | sed 's/\-/:/g') <(cut -c11- $DPath/named.query.log)
You can do the following:
cut -c 1-10 $DPath/named.query.log | sed -i 's/-/:/g'
The cut statemnt takes only the first 10 chars of each line in that file. The output of that should be piped in a file. As of now it will just output to your terminal