[mythtv-users] Sources and inputs

[Ma Begaj]

2008/7/9 Nick Morrott <knowledgejunkie at gmail.com>:
> 2008/7/8 Ma Begaj <derliebegott at gmail.com>:
>> 2008/7/8 Nick Morrott <knowledgejunkie at gmail.com>:
>
>>> * It is possible to use a single XMLTV file containing all channels
>>> you can receive on all sources, and run mythfilldatabase (with the
>>> --update switch) once for each source pointing to this file. It is
>>> also possible to use both XMLTV and EIT data on the same source to
>>> cover cases where the XMLTV grabber does not provide listings data for
>>> certain channels.
>>>
>>>
>>
>> it sound that using "--update" switch with  xmltv file should be
>> better than having two different files if you have same channels on
>> different sources?
>>
>> what will happen if you have a few same channels on different sources?
>> will mythfilldatabase get data for the same channel twice even if the
>> xmltv id is the same?
>
> When I last looked at mythfilldatabase a year or so ago, I'm sure
> there was no logic to apply XMLTV data to all matching XMLTVIDs on any
> configured source. Therefore, if you are running mythfilldatabase
> automatically the name of the XMLTV config file will be the name of
> the video source, and if you have >1 videosource, you need each config
> file to contain the channels available on that videosource, even if
> you have some common channels.
>
>> I am grabbing at the moment the data for one channel twice because it
>> is available on two different sources. I would like to able to
>> optimize grabbing, because xmltv is not the fastest one.
>
> If your grabber supports caching of retrieved data (some do, some do
> not), this will help to speed up processing as the grabber will not
> have to download the source data again. However, depending on how long
> the grabber takes to process the listings data and create the XMLTV
> output, caching may or may not give a noticeable speed increase.
>
> Alternatively, you can create a single XMLTV config file containing
> all channels you want to grab, and run the grabber manually - this is
> what I do. Then run MFDB with the --update and --sourceid flags to
> import the listings data for each videosource you have configured.
>
> I will try and do some speed tests to determine whether running MFDB
> automatically with a config for each videosource is any quicker than
> running MFDB manually with a single config file containing all
> channels. The grabber I use (tv_grab_uk_rt) supports caching, and
> because the XMLTV data is passed from the grabber to MFDB via STDOUT
> rather than via an intermediate file when running MFDB automatically,
> it may turn out to be more efficient overall to switch from running
> manually.
>

If someone has same channels on 2-3 different sources, the fastest way could
be to use config file without these channels at all. Then the I could
run the grabber
only for these channels and insert the same xml file manually for each source.

That could prevent getting the same data 2-3 times although it is not
a nice clean
solution. It should be handled by mythfilldatabase.

Do you know, is there some technical  reason why mythfilldatabase have "source"
flag? Couldn't mythfilldatabse just ignore "source" flag and only use
xmltv-id for identifying
it, because xmltv-id is/should be unique. It sounds to me that
xmltv-id from channels table
should be the foreign key in the programs table.