Sign up ×
Unix & Linux Stack Exchange is a question and answer site for users of Linux, FreeBSD and other Un*x-like operating systems. It's 100% free, no registration required.

I would like to know how I can get the value of a node with the following paths:

config/global/resources/default_setup/connection/host
config/global/resources/default_setup/connection/username
config/global/resources/default_setup/connection/password
config/global/resources/default_setup/connection/dbname

from the following XML:

<?xml version="1.0"?>
<config>
    <global>
        <install>
            <date><![CDATA[Tue, 11 Dec 2012 12:31:25 +0000]]></date>
        </install>
        <crypt>
            <key><![CDATA[70e75d7969b900b696785f2f81ecb430]]></key>
        </crypt>
        <disable_local_modules>false</disable_local_modules>
        <resources>
            <db>
                <table_prefix><![CDATA[]]></table_prefix>
            </db>
            <default_setup>
                <connection>
                    <host><![CDATA[localhost]]></host>
                    <username><![CDATA[root]]></username>
                    <password><![CDATA[pass123]]></password>
                    <dbname><![CDATA[testdb]]></dbname>
                    <initStatements><![CDATA[SET NAMES utf8]]></initStatements>
                    <model><![CDATA[mysql4]]></model>
                    <type><![CDATA[pdo_mysql]]></type>
                    <pdoType><![CDATA[]]></pdoType>
                    <active>1</active>
                </connection>
            </default_setup>
        </resources>
        <session_save><![CDATA[files]]></session_save>
    </global>
    <admin>
        <routers>
            <adminhtml>
                <args>
                    <frontName><![CDATA[admin]]></frontName>
                </args>
            </adminhtml>
        </routers>
    </admin>
</config>

Also I want to assign that value to the variable for further use. Let me know your idea.

share|improve this question
4  
Do not ever use bash to parse structured trees of arbitrary data. Use a real XML parser. I recommend XMLStarlet. –  Chris Down Jul 17 '13 at 15:16

5 Answers 5

up vote 8 down vote accepted

Using bash and xmllint (as given by the tags):

xmllint --version  #  xmllint: using libxml version 20703

# Note: Newer versions of libxml / xmllint have a --xpath option which 
# makes it possible to use xpath expressions directly as arguments. 
# --xpath also enables precise output in contrast to the --shell & sed approaches below.
#xmllint --help 2>&1 | grep -i 'xpath'

{
# the given XML is in file.xml
host="$(echo "cat /config/global/resources/default_setup/connection/host/text()" | xmllint --nocdata --shell file.xml | sed '1d;$d')"
username="$(echo "cat /config/global/resources/default_setup/connection/username/text()" | xmllint --nocdata --shell file.xml | sed '1d;$d')"
password="$(echo "cat /config/global/resources/default_setup/connection/password/text()" | xmllint --nocdata --shell file.xml | sed '1d;$d')"
dbname="$(echo "cat /config/global/resources/default_setup/connection/dbname/text()" | xmllint --nocdata --shell file.xml | sed '1d;$d')"
printf '%s\n' "host: $host" "username: $username" "password: $password" "dbname: $dbname"
}

# output
# host: localhost
# username: root
# password: pass123
# dbname: testdb

In case there is just an XML string and the use of a temporary file is to be avoided, file descriptors are the way to go with xmllint (which is given /dev/fd/3 as a file argument here):

set +H
{
xmlstr='<?xml version="1.0"?>
<config>
    <global>
        <install>
            <date><![CDATA[Tue, 11 Dec 2012 12:31:25 +0000]]></date>
        </install>
        <crypt>
            <key><![CDATA[70e75d7969b900b696785f2f81ecb430]]></key>
        </crypt>
        <disable_local_modules>false</disable_local_modules>
        <resources>
            <db>
                <table_prefix><![CDATA[]]></table_prefix>
            </db>
            <default_setup>
                <connection>
                    <host><![CDATA[localhost]]></host>
                    <username><![CDATA[root]]></username>
                    <password><![CDATA[pass123]]></password>
                    <dbname><![CDATA[testdb]]></dbname>
                    <initStatements><![CDATA[SET NAMES utf8]]></initStatements>
                    <model><![CDATA[mysql4]]></model>
                    <type><![CDATA[pdo_mysql]]></type>
                    <pdoType><![CDATA[]]></pdoType>
                    <active>1</active>
                </connection>
            </default_setup>
        </resources>
        <session_save><![CDATA[files]]></session_save>
    </global>
    <admin>
        <routers>
            <adminhtml>
                <args>
                    <frontName><![CDATA[admin]]></frontName>
                </args>
            </adminhtml>
        </routers>
    </admin>
</config>
'

# exec issue
#exec 3<&- 3<<<"$xmlstr"
#exec 3<&- 3< <(printf '%s' "$xmlstr")
exec 3<&- 3<<EOF
$(printf '%s' "$xmlstr")
EOF

{ read -r host; read -r username; read -r password; read -r dbname; } < <(
       echo "cat /config/global/resources/default_setup/connection/*[self::host or self::username or self::password or self::dbname]/text()" | 
          xmllint --nocdata --shell /dev/fd/3 | 
          sed -e '1d;$d' -e '/^ *--* *$/d'
       )

printf '%s\n' "host: $host" "username: $username" "password: $password" "dbname: $dbname"

exec 3<&-
}
set -H


# output
# host: localhost
# username: root
# password: pass123
# dbname: testdb
share|improve this answer
    
Man page: xmlsoft.org/xmllint.html –  Jason Pyeron Sep 21 '14 at 15:27

The following works when run against your test data:

{ read -r host; read -r username; read -r password; read -r dbname; } \
  < <(xmlstarlet sel -t -m /config/global/resources/default_setup/connection \
      -v ./host -n \
      -v ./username -n \
      -v ./password -n \
      -v ./dbname -n)

This puts the content into variables host, username, password and dbname.

share|improve this answer
    
xmlstarlet: command not found, so this command is not useful to me :( –  MagePsycho Jul 17 '13 at 15:31
    
@MagePsycho bash does not have any built-in support for XML parsing. You either need to have a tool that does (xmlstarlet, xsltproc, a modern Python, etc), or you can't parse XML correctly. –  Charles Duffy Jul 17 '13 at 15:48
    
@CharlesDuffy is there a way to get the value may be using regex pattern or else? –  MagePsycho Jul 17 '13 at 15:51
4  
@MagePsycho you can just install xmlstarlet. In any case, you should never use regular expressions to parse (X)HTML. –  terdon Jul 17 '13 at 15:57
1  
@MagePsycho I was about to post the same link terdon already did. In short: No. –  Charles Duffy Jul 17 '13 at 16:05

Although there are a lot of answers already, I'll chime in with xml2.

$ xml2 < test.xml
/config/global/install/date=Tue, 11 Dec 2012 12:31:25 +0000
/config/global/crypt/key=70e75d7969b900b696785f2f81ecb430
/config/global/disable_local_modules=false
/config/global/resources/db/table_prefix
/config/global/resources/default_setup/connection/host=localhost
/config/global/resources/default_setup/connection/username=root
/config/global/resources/default_setup/connection/password=pass123
/config/global/resources/default_setup/connection/dbname=testdb
/config/global/resources/default_setup/connection/initStatements=SET NAMES utf8
/config/global/resources/default_setup/connection/model=mysql4
/config/global/resources/default_setup/connection/type=pdo_mysql
/config/global/resources/default_setup/connection/pdoType
/config/global/resources/default_setup/connection/active=1
/config/global/session_save=files
/config/admin/routers/adminhtml/args/frontName=admin

With a little magic you can even set those as variables directly:

$ eval $(xml2 < test.xml | tr '/, ' '___' | grep =)
$ echo $_config_global_resources_default_setup_connection_host          
localhost
share|improve this answer
    
May work but unfortunately xml2 command is not found. –  MagePsycho Jul 19 '13 at 8:53
    
So install it with your package manager. –  bahamat Jul 19 '13 at 20:14

You can make use of php command line interface coding in bash scripts to handle several complex scripts that actually span over multiple lines of coding. First, try to make your solution using PHP scripts, and then later on pass the parameters using CLI mode. Thus, you can get control over superb usages of XML parsers.

The environment seems that you can use PHP in client mode via ssh/shell access.

php -f yourxmlparser.php

Now, do all the things within your php file. Make use of command line parameters it can take.

You can even assign that return values to Shell environment to continue rest of your shell scripts.

And the other way is to use |grep option to match your required value within the xml file, if you are pretty sure of the structure of your xml file that does not change over time.

share|improve this answer

A pure bash function, just for the unfortunate case when you are not allowed to install anything appropriate. This may, and probably will, fail on more complicated XML:

function xmlpath()
{
  local expr="${1//\// }"
  local path=()
  local chunk tag data

  while IFS='' read -r -d '<' chunk; do
    IFS='>' read -r tag data <<< "$chunk"

    case "$tag" in
      '?'*) ;;
      '!–-'*) ;;
      '![CDATA['*) data="${tag:8:${#tag}-10}" ;;
      ?*'/') ;;
      '/'?*) unset path[${#path[@]}-1] ;;
      ?*) path+=("$tag") ;;
    esac

    [[ "${path[@]}" == "$expr" ]] && echo "$data"
  done
}

Usage:

bash-4.1$ xmlpath 'config/global/resources/default_setup/connection/host' < MagePsycho.xml
localhost

Known issues:

  • slow
  • searches only by tag names
  • no character entity decoding
share|improve this answer

Your Answer

 
discard

By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.