[mythtv-users] Duplicate matching methods - make program ID an explicit choice

[Michael T. Dean]

On 05/19/2016 02:45 PM, John Veness wrote:
> On May 19, 2016 7:21:52 PM GMT+01:00, "Michael T. Dean" 
> <mtdean at thirdcontact.com> wrote:
> >> I haven't checked the raw output of the uk_rt grabber or the new SD
> >JSON one, but if it turns out to be the case that they both use
> >xmltv_ns season/episode numbering, and if program titles are the same,
> >maybe the autogenerated program IDs would match too? Does anyone else
> >know?
> >
> >That would be the case. However, I'd expect that the SD JSON one would
> >
> >use the dd_progid system.
>
> Someone on the xmltv-users mailing list has been posting snippets of 
> sample tv_grab_sd_json output for UK users, comparing with uk_rt. For 
> example, an episode of the series Blue Eyes has the following in uk_rt:
>
> <episode-num system="xmltv_ns">0.8/10.</episode-num>
>
> And the following in sd_json:
>
> <episode-num system="xmltv_ns">0.8.</episode-num>
> <episode-num system="dd_progid">EP023864610009</episode-num>
>
> I.e. it has both systems listed, and is missing the total number in 
> the xmltv_ns case. Would Myth use the dd_progid in this case, or an 
> autogenerated ID from the xmltv_ns text?

It would use the dd_progid first since that's meant to be the definitive 
identifier.

> I'd prefer it to use xmltv_ns, to make it more likely that duplicates 
> would be matched with my old recordings correctly.
>
> On the other hand, apparently some series in sd_json only output the 
> dd_progid, with no xmltv_ns (I.e. no season/episode numbering) which 
> is a shame. Those for sure will not be matched as duplicates with my 
> old uk_rt recordings.

You could always post-process your XMLTV data (but before 
mythfilldatabase uses the data) to remove some/all <episode-num 
system="dd_progid"> elements.

Mike