[mythtv-users] UK DVB-T Duplicates usually on Dave, Sky3 etc... how to remove?

Mike Holden mythtv at mikeholden.org
Fri Jun 6 13:52:35 UTC 2008


Nick Morrott wrote:
> On 06/06/2008, Mike Holden <mythtv at mikeholden.org> wrote:
>> There are 2 settings in the duplicate matching option, one is episode &
>>  description, the other is episode then description. If the program
>> doesn't
>>  have an episode title, make sure you use the latter of the 2 options
>> (i.e
>>  episode THEN description), otherwise all episodes match.
>>
>>  I'm not sure if this is a bug. My view is that if there is no episode
>>  title, then both options should match duplicates based on the
>> description
>>  only, but I didn't write the code :-)
>
> I find that since the introduction of 'subtitle then desc' I choose
> that option almost all of the time, just in case a subtitle is
> missing. I very rarely see repeats scheduled to record (usually due to
> a different description). The EIT data can also use the programme ID
> broadcast in the stream (if available) which offers yet more dupe
> detection.
>
> I mostly use the Radio Times data (no surprise there...) which I find
> very consistent, with RT-reviewed programmes having the reviewer's
> write-up the first time they're shown, later showings having the stock
> description. Non-reviewed series tend to have consistent
> episode/description information each time they're in the listings.
>
> The quality of data from different providers across the world will be
> different, so I think the ability to choose which detection strategy
> to use and have it applied by default to all future schedules might be
> a good thing. Therefore, if no-one beats me to it, I'll be adding a
> mythfrontend config option to allow a user to choose the default
> dupe-detection policy

Yep, I like that plan. I too use RT, but haven't yet got round to setting
up EIT on other channels - still building my new system.

I would much prefer to use episode then description as the default, as it
seems to cover listings "weirdness" best (missing ep title being the prime
one), and also works the way I would expect for programs that have no ep
title.

Back to my earlier question, do you consider it a bug that for a program
with no episode title, the 2 options behave differently? If there is no
episode name, then "&" and "then" __should__ both be equivalent, but it
doesn't work that way. If I had to guess, being from an Oracle database
background, NULL cannot be compared in the way you would expect. Here's an
example from a database in front of me:

SQL> select * from dual;

D
-
X

SQL> select * from dual where null = null;

no rows selected

"DUAL" is a table with a single row in it, as you can see in the first
query. However selecting "where null = null" rpoduces no rows, because you
can't use comparative operators like =, != <> etc with null. All you can
do is "where <item> IS NULL". In Oracle, an empty string is equivalent to
null.

The following 2 queries illustrate this:

SQL> select * from dual where '' is null;

D
-
X

SQL>  select * from dual where '' = '';

no rows selected

I don't know if mysql works this way as well, but it would explain the
issue we are seeing if the code is testing for equality between 2 empty
string values if it does work the same way as Oracle.
-- 
Mike Holden

http://www.by-ang.com - the place to shop for all manner of hand crafted
items, including Jewellery, Greetings Cards and Gifts





More information about the mythtv-users mailing list