How do I extract only top level elements in xml tree response in Linux shell script?

Question

Below is the shell script which displays the list of names and manager_add

name=($(grep -oP '(?<=name>)[^<]+' <<< "$vsppProxy_res"))

for i in ${!name[*]}
do
  echo "$i" "${name[$i]}"
  done

managers=($(grep -oP '(?<=manager_addr>)[^<]+' <<< "$vsppProxy_res"))

for i in ${!managers[*]}
do
  echo "$i" "${managers[$i]}"
  done

The out put so far is:

0 name0 1 name1 2 name2 3 name3 ........................ ........................ ........................ ........................ 0 manager_add0 1 manager_add1 2 manager_add2 ........................ ........................

The xml response contains 3 top level elements like, <elem id="0">and again each top level element contains sub elements like <elem id="0"> depending on the size. But, the requirement is need to extract only the top level element <elem id="0">

The sample xml looks like below :

<X>
<regs>
<elem id="0">
<id>1</id>
<name>name0</name>
<warn>1</warn>
<manager_addr>manager_addr0</manager_addr>
<warn_desc>
<size>14</size>
<elem id="0">
<sev>2</sev>
<description>description</description>
<warning_id>1</warning_id>
<deployment_id>1</deployment_id>
<context_id>00</context_id>
<num_of_occurrences>1</num_of_occurrences>
<deployment_name>prod1</deployment_name>
</elem>
<elem id="1">
<sev>2</sev>
<description>description</description>
<warning_id>1</warning_id>
<deployment_id>1</deployment_id>
<context_id>00</context_id>
<num_of_occurrences>1</num_of_occurrences>
<deployment_name>prod1</deployment_name>
</elem>
</warn_desc>
</elem>
<elem id="1">
<id>2</id>
<name>name1</name>
<warn>1</warn>
<manager_addr>manager_addr1</manager_addr>
<warn_desc>
<size>1</size>
<elem id="0">
<sev>3</sev>
<description>description</description>
<warning_id>2</warning_id>
<context_id>00</context_id>
<num_of_occurrences>1</num_of_occurrences>
</elem>
</warn_desc>
</elem>
</regs>
</X>

The expected output should be :

Manager: name0 manager_add0 ........................ ........................ ........................

Please provide your suggestions. Thank you.

Use a tool that actually understands xml. Maybe a xslt stylesheet? — Shawn, Jan 22 '20 at 08:20
A complete well-formed XML input sample would be useful, btw. — Shawn, Jan 22 '20 at 08:24
And your expected output based on that input? (Also note well-formed part - the new xml has issues. Opens with ``, closes with ``?) — Shawn, Jan 22 '20 at 08:46
I has just edited the xml response, we need only the top level elements in between ... like — R venkata, Jan 22 '20 at 09:24
this might be helpfull. [stackoverflow.com/questions/893585/how-to-parse-xml-in-bash](https://stackoverflow.com/questions/893585/how-to-parse-xml-in-bash) — Nitish, Jan 22 '20 at 09:29

score 1 · Answer 1 · answered Jan 22 '20 at 09:29

Trying to use regular expressions to parse XML is a Bad Idea. You should be using XML-aware tools. One obvious choice for transforming an XML document is XSLT.

Consider the following shell script, which applies a XSLT stylesheet to the file passed to it on the command line:

#!/bin/sh

# Use a temporary file for the XSLT stylesheet
stylesheet=$(mktemp)
cat >"$stylesheet" <<'EOF'
<?xml version="1.0" encoding="UTF-8"?>
<xsl:stylesheet xmlns:xsl="http://www.w3.org/1999/XSL/Transform" version="1.0">
  <xsl:output method="text" encoding="UTF-8" />
  <xsl:variable name="newline"><xsl:text>
</xsl:text></xsl:variable>
  <xsl:template match="/X">
    <xsl:for-each select="regs/elem">
      <xsl:value-of
          select="concat('Manager: ', name, ' ', manager_addr, $newline)" />
    </xsl:for-each>
  </xsl:template>
</xsl:stylesheet>
EOF

# Adjust as needed for whichever xslt processor you want to use
xsltproc "$stylesheet" "$1"
# xalan -xsl "$stylesheet" -in "$1"
# xmlstarlet tr "$stylesheet" "$1"

rm -f "$stylesheet"

Usage:

$ ./transform.sh input.xml
Manager: name0 manager_addr0
Manager: name1 manager_addr1

How do I extract only top level elements in xml tree response in Linux shell script?

1 Answers1