@ginjaninja said:
append to response alternative titles on query
https://api.themoviedb.org/3/search/movie?api_key=THE_KEY&append_to_response=alternative_titles&query=Feedback&primary_release_year=2019&page=1

I have read elsewhere on a 7 year old post that append to response is not available when doing queries on movies. is that still the case?

Yes, it still is.

It seems a little bit problematic for an app trying to find matches on tmdb, because as in the above example the correct 1st search result is based on a match by alternative title (i believe) but from the initial response there is no way to know that.

Do not. It is of the titles in Germany and Spain.

I could request the alternative titles for every object returned in the first query but thats quite a load on the api / performance penalty.
my goal is to provide the movies title and year and find the right tmdbid for that movie.
is there a recommended approach to this?

There are many titles that have common words.
Stone, Star, War, Feedback etc...
Thinking about my search experiences, I would say that at most 20% of searches only return one movie.
Therefore, I suggest that if more than one movie appears in the survey response using your APP,
open a screen with the list of movies in the response.
If possible showing the Poster.
And choose the desired movie.

I see no other solution.

Hi tiaco2, thanks for help. alternative titles for 577889 are

iso_3166_1 title       type
---------- -----       ----
BR         Vingança FM
GB         Feedback

I am fairly sure the api is using alternative_titles in a normal movie query

https://api.themoviedb.org/3/search/movie?api_key=THE_KEY&query=Feedback&primary_release_year=2019&page=1

so a gb user might well have named the movie

Feedback (2019)

my app (which is headless for this particular gettmdbid function) is trying to sift through edge cases where its unclear what is the best match and going by the response above the logic could decide the 2nd object is the correct one, but it is in fact the first. often it is not possible to trust the 1st response because the user is mistaken on the release year so a wordmatch / string proximity function can disambiguate/improve confidence, but that needs the match title to work on.

One solution might be to support the appending of alternative titles in the movie query response.

@ginjaninja said:

Hi tiaco2, thanks for help. alternative titles for 577889 are

iso_3166_1 title       type
---------- -----       ----
BR         Vingança FM
GB         Feedback

I am fairly sure the api is using alternative_titles in a normal movie query

I think I didn't understand very well. Something is always lost in translation.

Yes, use. All titles, official and alternative.
I just wanted to say that it wasn't just because of the alternatives, that there are officers with the word.

https://api.themoviedb.org/3/search/movie?api_key=THE_KEY&query=Feedback&primary_release_year=2019&page=1

so a gb user might well have named the movie

Feedback (2019)

my app (which is headless for this particular gettmdbid function) is trying to sift through edge cases where its unclear what is the best match and going by the response above the logic could decide the 2nd object is the correct one, but it is in fact the first. often it is not possible to trust the 1st response because the user is mistaken on the release year so a wordmatch / string proximity function can disambiguate/improve confidence, but that needs the match title to work on.

This is where I least understood.
Just a guess of mine, as I don't know the answers sort order rules.
It seems to me that the tendency is to put the most recent, or most famous, or most popular ones first.
So it is not possible to believe that the first in the list is necessarily the correct one.
Especially when looking for older movies or non-famous movies.

One solution might be to support the appending of alternative titles in the movie query response.

See if these 2 proposals clarify anything.
https://trello.com/c/FhpxaRAx/132-support-searching-for-exact-matches
https://trello.com/c/0ygbkI4F/41-implement-search-that-factors-in-querying-language

thanks the support searching for exact matches might be interesting. For now im getting the alternative titles and applying a weighting system to compare factors and building a confidence score out of that. I can get the right id 3497 times out of 3500, and as long a 3 incorrect have a confidence score to encourage the users to check, then thats good enough. The factors im using are

• word match
• string distance, scales with discrepancy
• 10x/100x swings in vote count
• native language and preferred languages
• year discrepancy, weight scales with discrepancy

if the api is already doing the work to check queries against alternative titles, then my feature request still stands to include alternative titles in the api response (or allow for it as an option)