I have an XML-like text file, which cannot be parsed with an XML parser due to XML violations:
<note> <to>Tove</to> <from>Jani</from> <heading>Reminder</heading> <body>Don't forget me this weekend!</body> </note> I just want to cut the word after n th tag open < in a file. The file should be in XML-grammar, which means lines can vary.
My expected output would be,
1 - note 2 - to 3 - /to 4 - from 5 - /from 6 - heading 7 - /heading 8 - body 9 - /body 10 - /note 
$ awk -F"[<>]" '{for(i=2;i<=NF;i+=2){print ++j" - "$i}}' input.xml 1 - note 2 - to 3 - /to 4 - from 5 - /from 6 - heading 7 - /heading 8 - body 9 - /body 10 - /note Note: This answer was written before the user explained that the XML was not well formed. I'm leaving it here as it may possibly help others.
XMLStarlet is able to produce the element structure of XML documents:
$ xml el file.xml note note/to note/from note/heading note/body This is different from your expected output, but may be enough for what you want to achieve.
It is also able to convert the XML to PYX, which shows the opening and closing tags on separate lines:
$ xml pyx file.xml (note -\n (to -Tove )to -\n (from -Jani )from -\n (heading -Reminder )heading -\n (body -Don't forget me this weekend! )body -\n )note From this, it's easy to get exactly the output you are after:
$ xml pyx file.xml | sed -n -e 's/^(//p' -e 's/^)/\//p'| nl 1 note 2 to 3 /to 4 from 5 /from 6 heading 7 /heading 8 body 9 /body 10 /note The sed instructions removes lines not starting with either ( or ) and replaces these characters according to how you specified it in the question. The nl utility puts line number on lines.
XMLStarlet is sometimes installed as xmlstarlet rather than xml.
grep + awk solution:
grep -Eo '<[^<>]+>' input.xml | awk '{ gsub(/[<>]/,""); printf "%-3s - %s\n", NR, $0 }' The output:
1 - note 2 - to 3 - /to 4 - from 5 - /from 6 - heading 7 - /heading 8 - body 9 - /body 10 - /note Or with single GNU awk command:
awk -v FPAT='</?[^<>]+>' '{ for(i=1;i<=NF;i++) printf "%-3s - %s\n", ++c, $i }' input.xml here is an quite easy method to answer your question on extracting openning tags... but your example ask also for closing ones .... this seems nosense because a closed one is open of course.... do you really need also closed ones but if you wanna control xml format but the use a tool like xmllint ....
bash-4.4$ cat > toto <note> <to>Tove</to> <from>Jani</from> <heading>Reminder</heading> <body>Don't forget me this weekend!</body> </note> bash-4.4$ awk '{ match($0,/<\/.*>/); b=substr($0,RSTART,RLENGTH); if(b) {a[++i]=b} } END{ {for(k in a) {c[a[k]]=k} } {for(u in c) {gsub(/\//,X,u);print u} } }' toto | sed 's/</- /;s/>//' | cat -n 1 - body 2 - note 3 - to 4 - heading 5 - from bash-4.4$ rm toto or to keep all & using sed only for fun :
bash-4.4$ sed -e 's/>\(.*\)</></;s/>/\n/g;s/</- /g' toto | sed '/^$/ d' | cat -n 1 - note 2 - to 3 - /to 4 - from 5 - /from 6 - heading 7 - /heading 8 - body 9 - /body 10 - /note 11 bash-4.4$ Here's an XQuery solution just in case you want something that works on ANY XML, even awkward XML containing comments, DTDs, self-closing elements, etc.
declare function local:f($e) { $e / (name(), local:f(*), ('/' || name())) }; for $tag at $p in local:f(/*) return ($p || ' - ' || $tag || '
') 
<note> <to><to_1><to_1_1>to_child1_1</to_1_1><to_1_2>to_child2_2</to_1_2></to_1><to_2>to_child2</to_2></to> <from>Jani</from><heading>Reminder</heading></note>What should be the 5th node in such case?