complex sed multiline match and replace

Question

<Placemark id="051314">
<name>HI Hostel</name>
<description><![CDATA[<div style="color: #404040;font-size: 12px"><a "#book"style="color:#295181;font-size: 12px" target="_top" href="http://www.hihostels.com/dba/hostel051314.de.htm?himap=Y#book" >Girona - Equity Point Girona</a><img style="margin: 5px 0px 5px 0px; border-color:#909090; padding:2px; display:block; clear:both;" src="http://www.hihostels.com/pics/ES/051314_pic_main.jpeg" width="96" height="72" border="1">Plaça Catalunya, 23<br>Girona<br>17002<br><b>Spanien</b><br><div style="margin-top:3px;"><img style="vertical-align:middle;margin-right:5px;" src="http://www.hihostels.com/imgfront/pegsmall.png" /><a style="color:#295181;font-size: 12px;" href="http://www.hihostels.com/openSVwindow(41.981658,2.823057)">Street View</a></div></div> ]]></description>

My source files look like the one above (basically coming from http://www.hihostels.com/mapcoord/ES.en.kml ). I want to replace the (useless) name tag "HI Hostel" (always the same for every placemark) with the hostels real name. The real name appears in the description tag one line below, in the case above it would be "Girona - Equity Point Girona".

Any clever idea on how to do this? Thanks for reading.

Answer 1

Some like this? Using awk

awk -F, '/^<name>/ {next} /^<description/ {s=$0;gsub(/<[^>]*>/, ",");$0="<name>" $4 "</name>\n" s} 1' file
<Placemark id="051314">
<name>Girona - Equity Point Girona</name>
<description><![CDATA[<div style="color: #404040;font-size: 12px"><a "#book"style="color:#295181;font-size: 12px" target="_top" href="http://www.hihostels.com/dba/hostel051314.de.htm?himap=Y#book" >Girona - Equity Point Girona</a><img style="margin: 5px 0px 5px 0px; border-color:#909090; padding:2px; display:block; clear:both;" src="http://www.hihostels.com/pics/ES/051314_pic_main.jpeg" width="96" height="72" border="1">Plaça Catalunya, 23<br>Girona<br>17002<br><b>Spanien</b><br><div style="margin-top:3px;"><img style="vertical-align:middle;margin-right:5px;" src="http://www.hihostels.com/imgfront/pegsmall.png" /><a style="color:#295181;font-size: 12px;" href="http://www.hihostels.com/openSVwindow(41.981658,2.823057)">Street View</a></div></div> ]]></description>

This may also work:

awk -F"<|>" '/^<name>/ {next} /^<description/ {$0="<name>" $8 "</name>\n" $0} 1' file