Sed: how to replace nextline \n symbol in text files?

Question

I need to fix an error and to replace the second tag </time> with </tags> in an XML file with the following structure:

<time>20260664</time>
<tags>substancesummit ss</time>
<geo>asdsadsa</geo>
<time>20260664</time>
<tags>substancesummit ss</time>
<geo>asdsadsa</geo>

I'm trying to do it using sed and since I have 2 </time> closing tag per item, my idea is to replace </time><geo> with </tags><geo>.

However there is a next line symbol in between, so I'm using \n but it doesn't work:

sed 's/time>\n<geo>/tags>\n<geo>/g' old.xml > new.xml

Any help?

@Rafe Kettler It doesn't replace the content. The output is the same of the original. — user680406, Apr 23 '11 at 14:58
possible duplicate of [SED: How can I replace a newline (\n)?](http://stackoverflow.com/questions/1251999/sed-how-can-i-replace-a-newline-n) — dmckee --- ex-moderator kitten, Apr 23 '11 at 15:42

anubhava · Accepted Answer · 2011-04-25T03:49:39.707

You can do that in 1 single sed command like this:

sed '/<\/time>/I{n;:A;N;h;/<geo>/I!{H;bA};/<geo>/I{g;s/<\/time>/<\/tags>/i}}' file.txt

Testing

If your input file.txt is like this:

<time>20260664</time>
<tags>substancesummit ss
</time>

<Geo>asdsadsa</geo>
<time>30260664</time>
<tags>substancesummit st</timE>
<geo>bsdsadsa</geo>

Then output of above command will be:

<time>20260664</time>
<tags>substancesummit ss
</tags>

<Geo>asdsadsa</geo>
<time>30260664</time>
<tags>substancesummit st</tags>
<geo>bsdsadsa</geo>

It covers multiple new line characters (\r or \n) in any combination between </time> and <geo>

PS: Above sed command is doing ignore care search/replace, in case you don't want that then just remove I flag from sed command or just let me know.

score 1 · Answer 2 · answered Apr 23 '11 at 16:04

1

Use this:

$ sed -n '1h; 1!H; $ {g; s/<\/time>\n<geo>/<\/tags>\n<geo>/g; p;}' file

answered Apr 23 '11 at 16:04

sahaj

822
5
17

score 1 · Answer 3 · answered Apr 23 '11 at 16:46

If there is a character that you definitely don't use in the file, try to replace \n with it, do your sed work and replace back. tr works really well for that

cat old.txt | tr '\n' '#' | sed 's/time>#<geo>/tags>#<geo>/g' | tr '#' '\n' > new.txt

I use # as replace character.

score 0 · Answer 4 · answered Apr 23 '11 at 15:02

sed usually edits lines, and it is a bit harder to make it understand multiple lines at once, as you are. Instead, how about fixing up the broken lines more directly, with something like this:

/<tags>/ s@</time>@</tags>@

This will replace </time> with </tags> only on lines which also contain <tags>. Note that I used @ instead of / as the delimiter for the substitution command, to avoid a need to escape the slashes in the XML we're trying to replace.

score 0 · Answer 5 · answered Apr 23 '11 at 15:14

you can useawk instead

$ awk -vRS="</geo>" '{gsub(/<\/time>.<geo>/,"</tags>\n<geo>")}1' ORS="</geo>" file
<time>20260664</time>
<tags>substancesummit ss</tags>
<geo>asdsadsa</geo>
<time>20260664</time>
<tags>substancesummit ss</tags>
<geo>asdsadsa</geo>

First, I can see that </geo> ends each block, so make this the record separator. After that, substitute what is required. Lastly, put </geo> back as the output record separator (ORS).

score 0 · Answer 6 · answered Apr 23 '11 at 16:37

Why don't you sidestep the issue with trying to match the linebreaks and instead try matching the line with the opening <tags> tag and the content after it up to the (non-)matching </time> tag? Like

# untested, written from scratch
sed 's/<tags>(.*)<\/time>/<tags>\1<\/tags>/g' infile > outfile

score 0 · Answer 7 · answered Apr 23 '11 at 16:43

sed -e 's,<\([^>]*\)>\([^<]*\)</[^>]*>,<\1>\2</\1>,g' tags.xml

This replaces in the same line an

(opening tag)(content)(closing tag)

with

(opening tag)(content)(closing tag)

but the closing tag is always the same tag, as the opening tag.

It can fail if more than one tag-pair is found in the file.

In Detail, it searches for something, starting with '<', followed by the name of the tag without closing it '>', followed by the content, which is everything until '<'.

Sed: how to replace nextline \n symbol in text files?

7 Answers7

Testing