Find and conditionally edit text in an XML file - bash

Question

I have a (XML-)file that has the following content:
<class>OverAll</class>
<char>
<rank> 1</rank>
<name> yyy</name>
<level> 9</level>
<experience>53842</experience>
<class>xxx</class>
</char>
<char>
<rank> 2</rank>
<name>aaa</name>
<level> 9</level>
<experience>53074</experience>
<class>zzz</class>
</char>
..and so on. I want to extract the number between the <experience> </experience> lines and replace it with a modified version of the number I found between the tag. For example, the file should look like this after the script:
<class>OverAll</class>
<char>
<rank> 1</rank>
<name> yyy</name>
<level> 9</level>
<experience>53.842</experience>
<class>xxx</class>
</char>
<char>
<rank> 2</rank>
<name>aaa</name>
<level> 9</level>
<experience>53.074</experience>
<class>zzz</class>
</char>
(I want to add a thousands separator, and values above 1 Million is required. So 2 thousand Separators :)
I am able to find and replace the number, but I dont know how to use the input number and modify it and add it back to the line.
Perhaps someone can help here?
Thank you very much :)

Answer 1

A one-liner sed can do it, assuming the last three digits are always decimal:
sed -zE 's#([[:digit:]]{7,})([[:digit:]]{1})[[:space:]]*(</experience[[:space:]]*>)#\1.\2\3#g;s#([[:digit:]]{3})[[:space:]]*(</experience[[:space:]]*>)#.\1\2#g'
sed parameters breakdown:
-zE
-z or --null-data: Separate lines by NULL characters to allow pattern matching across lines, because spaces, tabs and newlines are allowed by the XML syntax before the > bracket of a tag.
-E or --regexp-extended: Use extended regular expressions in the script (for portability use POSIX -E).
s#([[:digit:]]{7,})([[:digit:]]{1})[[:space:]]*(</experience[[:space:]]*>)#\1.\2\3#g:
Insert a decimal point before the last digit, to experience numbers containing seven plus one (eight) or more digits (Million or more with an extra decimal digit).
s#([[:digit:]]{3})[[:space:]]*(</experience[[:space:]]*>)#.\1\2#g:
Insert a decimal point before the last three digits, to experience numbers ending with three digits (automatically excludes the Millions experiences already processed by previous sed command.
Now keep in mind that it is not parsing the XML either, because it will replace numbers in the <experience> tag anywhere in the XML tree.
Regular expressions are not meant to parse markup languages. There are better, more efficient and dedicated tools to manipulate XML with XSLT/XPATH like saxon, xsltproc, xmllint...
Using proper XML processing with xsltproc:
decimal-experience.xsl
<?xml version="1.0"?>
<xsl:stylesheet version="1.0" xmlns:xsl="http://www.w3.org/1999/XSL/Transform">
<!-- Cosmetic sugar to have the xml declaration header and indent -->
<xsl:output omit-xml-declaration="no" indent="yes"/>
<!-- Cosmetic sugar to remove unneeded spaces in elements -->
<xsl:strip-space elements="*"/>
<!-- Copy all the nodes as-is from the source xml -->
<xsl:template match="*">
<xsl:copy>
<xsl:apply-templates select="node()"/>
</xsl:copy>
</xsl:template>
<!-- Process the content of the experience tag within the char tag -->
<xsl:template match="char/experience/">
<!-- If the experience is not already in decimal form -->
<xsl:if test="not(contains(., '.'))">
<xsl:choose>
<!-- When the experience is less than a Million -->
<xsl:when test=". < 9999999">
<!-- The last three digits are decimals -->
<xsl:value-of select="format-number(. div 1000, '0.000')"/>
</xsl:when>
<!-- Otherwise the experience is a Million or more -->
<xsl:otherwise>
<!-- The last digit is decimal -->
<xsl:value-of select="format-number(. div 10, '0.0')"/>
</xsl:otherwise>
</xsl:choose>
</xsl:if>
</xsl:template>
</xsl:stylesheet>
Running the XSLT transformation above:
xsltproc decimal-experience.xsl characters.xml
Example output:
I created a valid fictive characters.xml with a span root tag, because your extract was invalid XML.
<?xml version="1.0"?>
<span>
<class>OverAll</class>
<char>
<rank> 1</rank>
<name> yyy</name>
<level> 9</level>
<experience>53.842</experience>
<class>xxx</class>
</char>
<char>
<rank> 2</rank>
<name>aaa</name>
<level> 9</level>
<experience>53.074</experience>
<class>zzz</class>
</char>
<char>
<rank> 3</rank>
<name>Million</name>
<level>42</level>
<experience>5585307.4</experience>
<class>zzz</class>
</char>
</span>