How to read a value from recursive xml attribute in Unix using sed/awk/grep only

Question

I have config.xml. Here I need to retrieve the value of the attribute from the xpath /domain/server/name

I can only use grep/sed/awk. Need Help

The content of the xml is below where I need to retrieve the Server Name only.

<domain>
    <server>
        <name>AdminServer</name>
        <port>1234</port>
    </server>
    <server>
        <name>M1Server</name>
        <port>5678</port>
    </server>
    <machine>
        <name>machine01</name>
    </machine>
    <machine>
        <name>machine02</name>
    </machine>
</domain>

The output should be :

AdminServer
M1Server

I tried to do,

sed -ne '/<\/name>/ { s/<[^>]*>(.*)<\/name>/\1/; p }' config.xml

sed -ne '/<\/name>/ { s/<[^>]*>\(.*\)<\/name>/\1/; p }' config.xml This command is returning the values from all the name atttributes AdminServer M1 machine01 machine02 — jkalyanc, Jul 20 '17 at 11:29
sed -ne '//,\||{ s/ <[^>]*>\(.*\)<\/server>/\1/; p }' config.xml . On using this command the result is AdminServer 1234 M1Server 5678 Atleast want to know if i can use this output and fetch the server name? — jkalyanc, Jul 20 '17 at 11:35
@RomanPerekhrest No we dont have python either. I am actually in a tricky situation — jkalyanc, Jul 20 '17 at 11:37
@jkalyanc: Can you print `awk --version` and same for `grep` and `sed`? — Inian, Jul 20 '17 at 11:44

Ed Morton · Answer 1 · 2017-07-20T12:30:25.723

sed is only for simple substitutions on individual lines, doing anything else with sed is strictly for mental exercise, not for real code. That's not what you are trying to do so you shouldn't even be considering sed. Just use awk:

$ awk -F'[<>]' 'p=="server" && $2=="name"{print $3} {p=$2}' file
AdminServer
M1Server

That will work with any awk on any UNIX box. If that's not all you need then edit your question to provide more truly representative sample input and expected output.

score 2 · Answer 2 · answered Jul 20 '17 at 11:54

2

Try this command. Name your xml and supply that file as an input.

awk '/<server>/,/<\/server>/' < name.xml | grep "name" | cut -d ">" -f2 | cut -d "<" -f1

OutPut:

AdminServer
M1Server

answered Jul 20 '17 at 11:54

Ajay2588

527
3
6

score 1 · Accepted Answer · answered Jul 20 '17 at 11:58

1

Based on your sample Input_file shown, could you please try following.

awk -F"[><]" '/<\/server>/{a="";next} /<server>/{a=1;next} a && /<name>/{print $3}'  Input_file

answered Jul 20 '17 at 11:58

RavinderSingh13

130,504
14
57
93

score 1 · Answer 4 · answered Jul 21 '17 at 06:10

sed -n '/<server>/{n;s/\s*<[^>]*>//gp}'

for example. for the first match

1. /<server>/
match the line that contains "<server>" got "     <server>"

2. n
the "n" command will go to next line. after executed "n" command got "        <name>AdminServer</name>"

3.s/\s*<[^>]*>//gp
replece all "\s*<[^>]*>" as "". then print the pattern space

type "info sed" for more sed command

score 0 · Answer 5 · answered Jul 20 '17 at 12:36

0

You can get the desired output with just sed:

sed -n 's:.*<name>\(.*\)</name>.*:\1:p' config.xml

answered Jul 20 '17 at 12:36

Benjamin Connelly

47
1
6

score 0 · Answer 6 · answered Jul 20 '17 at 13:25

I feel dirty parsing XML in awk.

The following finds the correct depth of entry with the right tag name. It does not verify the path, though it depends on the elements you specified. While this works on your example data, it makes certain ugly assumptions and it's not guaranteed to work elsewhere:

awk -F'[<>]' '$2~/^(domain|server|name)$/{n++} $1~/\// {n--} n==3&&$2=="name"{print $3}' input.xml

A better solution would be to parse the XML itself.

$ awk -F'[<>]' -v check="domain.server.name" '$2~/^[a-z]/ { path=path "." $2; closex="</"$2">" } $0~closex { sub(/\.[^.]$/,"",path) } substr(path,2)==check {print path " = " $3}' input.xml
.domain.server.name = AdminServer

Here it is split out for easier commenting.

$ awk -F'[<>]' -v check="domain.server.name" '
  # Split fields around pointy brackets. Supply a path to check.

  $2~/^[a-z]/ {         # If we see an open tag,
    path=path "." $2    # append the current tag to our path,
    closex="</"$2">"    # compose a close tag which we'll check later.
  }

  $0~closex {                   # If we see a close tag,
    sub(/\.[^.]$/,"",path)      # truncate the path.
  }

  substr(path,2)==check {       # If we match the given path,
    print path " = " $3         # print the result.
  }

' input.xml

Note that this solution barfs horribly if you feed it badly formatted XML. The recognition of tags could be improved, but may be sufficient if you have consistently formatted XML. It may barf horribly for other reasons too. Do not do this. Install the correct tools to parse XML properly.

How to read a value from recursive xml attribute in Unix using sed/awk/grep only

6 Answers6