xmltv import and 'clumps'

[nielm]

It seems that the import plugin does not handle 'clumps' of programs very well -- these are when multiple programs share the same timeslot, which happens sometimes with RT grabber, eg:

last night: Film: Thirteenth Floor, followed by Weatherview (http://www.radiotimes.com/jsp/frameset/index.html?src=http://www.radiotimes.com:80/ListingsServlet?event=10&channelId=94&programmeId=15846669&jspLocation=/jsp/prog_details.jsp&jspError=/jsp/popup_error.jsp)...

This shows up in the XML output as two programs with the same timeslot but with clumpidx=0/2 and 1/2:
<programme start="20040128232500 +0000"
stop="20040129010500 +0000"
channel="london.bbc1.bbc.co.uk"
clumpidx="0/2">
<title lang="en">The Thirteenth Floor</title>
</programme>
<programme start="20040128232500 +0000"
stop="20040129010500 +0000"
channel="london.bbc1.bbc.co.uk"
clumpidx="1/2">
<title lang="en">Weatherview</title>
</programme>


The XMLTV importer ignores this clumpidx and over-writes the first program with the second... Which meant I missed File 2004 on Wed, and 13th Floor last night :( -- they were both described as 'Weatherview!'.

A workaround is to use tv_grep to grep for non-clumped programs, and only the first program in each clump (which is usually the most important):

tv_grep --not --clumpidx "" -or --clumpidx "0/" data.xml > noclumps.xml

[nielm]

Any moderators around, would it be possible to make this a 'sticky note' -- thanks!

[StephaneM]

It seems that the import plugin does not handle 'clumps' of programs very well
Just a quick note, if you use my Import plugin, the whole "tv_grep" thing is not needed. The Importer will correctly handle the "clumps", that is : retrieving the first show in the clump, and adding the other shows in the description of the first one.

Regards,
Stéphane.