I'm trying to retrieve all videos for a channel, and some are not being returned by the api. I cannot find anything in the spec that indicates why some would not be in the result set.
The call I'm making is:
https://www.googleapis.com/youtube/v3/search?type=video&key=__key_here__&channelId=UCxS2lX7728bTnmK1t21bYlA&part=id,snippet&maxResults=50&order=title
[To test this you'll need your own api key]
The first page of results is missing at least one video. The one from 8-15-2018, titled I LEARNED HOW TO DO A NEW TRICK!, which is located here, is not in the result set, even though it falls within the date range, and the first 50 results.
Does anyone know if this is a known issue?
I have verified that if I add the q parameter, with the video id, it will retrieve it.
You may refer with this link.
The API call that you should make if you want to get the videos in a channel is a youtube.playlistItems.list() with the playlistId of the "uploads" playlist for the channel. (This usually stars with UU..., but that's an implementation detail that might change in the future.
Here's an example of the call in the API Explorer.
Getting the same data via a search operation isn't guaranteed to return everything; the search index isn't a replacement for the backend data that you can obtain via the youtube.playlistItems.list() call. It's very much the same point raised in this blog post (though the focus there is on v2).
To summarize, to get all the uploads from a channel, you need to get the items from the uploads playlist for the channel using playlistItems.list on that playlist's ID rather than calling search.list on the channel ID.
Related
I am trying to collect a large list of YouTube's most watched videos for a data science application. I tried to use the YouTube Data API with the following query:
https://www.googleapis.com/youtube/v3/search&order=viewCount&type=video®ionCode=US&key=API_KEY,
but it does not seem to give me the same video ideas as in this list: https://en.wikipedia.org/wiki/List_of_most-viewed_YouTube_videos
Could someone tell me how I should do it?
Using the Search.list API endpoint is one way to search for the most viewed videos pertaining to a given region.
There's another endpoint -- Videos.list -- that, when queried with chart set to mostPopular, gives back:
chart (string)
The chart parameter identifies the chart that you want to retrieve.
Acceptable values are:
mostPopular – Return the most popular videos for the specified content region and video category.
As per the specification of videoCategoryId:
By default, charts are not restricted to a particular category.
Therefore, you may safely invoke Videos.list with chart=mostPopular and regionCode=US, without passing to it any videoCategoryId.
These two are the only API means that accomplish the task described by your post.
So I'm using Home Assistant to launch an automation that retrieves the newest videoId that a channel has uploaded, so I can use my google home to play it(on a Roku TV), works fine, I am working on creating an automation that also does a GET request but for now, I am using the home assistant rest sensor that updates by performing a GET after a set number of seconds, for some reason though there are only 3 sensors polled every minute or so it seems they use around 100-500 quota(hitting my quota of 10,000 after only a few hours or less), I'm not sure if this is a home assistant problem or if I am not using the api correctly(I only need the videoId), ill link my url below:
https://www.googleapis.com/youtube/v3/search?key=API_KEY&part=id&order=date&maxResults=1
Expected 1-3 quota usage per GET, getting 100+ quota usage per GET.
Querying the Search Endpoint is more expensive than querying the PlaylistItems endpoint for the given user's uploads playlist. Depending on usage patterns, the default quotas may put rather tight limits on the number of calls an user is allowed to make on various endpoints of the API.
Adapting my answer to a different question, I suggest you to do the following instead: call PlaylistItems endpoint, passing to it as playlistId parameter the given channel's uploads playlist ID.
A given channel's uploads playlist ID is obtained upon querying the channel's own endpoint. The needed ID is to be found at .items.contentDetails.relatedPlaylists.uploads. Usually, an channel ID and its corresponding uploads playlist ID are related by s/^UC([0-9a-zA-Z_-]{22})$/UU\1/.
Note that you should query the Channels endpoint only once, then use the returned uploads playlist ID as many times as you need.
Also note that you may experiment using the fields parameter applied to your queries, as to get from the API partial resources only. However, I'm predicting that (I may well be wrong, since did not tested it) the cost of 3 points for querying PlaylistItems for its contentDetails object cannot be improved.
Here is a prototype URL:
https://www.googleapis.com/youtube/v3/playlistItems?key=APP_KEY&part=contentDetails&fields=items/contentDetails/videoId&maxResults=1&playlistId=PLAYLIST_ID
It seems that the YouTube API doesn't give updated results for mostPopular videos in my country since few days.
Example:
This request (https://www.googleapis.com/youtube/v3/videos?part=snippet,contentDetails&chart=mostpopular®ionCode=FR) doesn't give me the same videos results than the ones displayed directly on YouTube for the French most popular channel (https://www.youtube.com/channel/UCmzy72gDEpfXoFV9Xdtd0DQ). It seems that the results of this request is not updated since the 1th of february. Results was real time updated before this.
Does someone know if something is wrong with my API request, or if there are some issues with the YouTube API at this moment?
There is nothing wrong with your request. This is a known issue with the YouTube API reported here for Saudi Arabia (but also applicable to multiple regions), and another related issue here with regard to content from France.
Your best bet would be to follow up with the YouTube team on one of those defects, or potentially (and dangerously) scrape the YouTube site for the correct results.
Problem seems to be solved since 13th of february (maybe someone from Google have seen my post..)
YouTube Channels and chart=mostPopular parameter data are separate data entities, aka you will get different results. They may be related but there is no guarantee you will get the same data. To get the data that you want you may need to query for channel itself and its videos.
I got this information from the thread #Jal linked, there was an update by matthewc...#google.com a few days ago:
The most popular channel for Saudia
Arabia and
the mostPopular chart parameter in the video.list
call
are separate and distinct entities. If you'd like to get the content
of the most popular channel for Saudia
Arabia
please use the Data API video.list call to list the videos with the
channel ID (in this case "UCWY-_j1MCth6yf24m58Bh_Q") by setting the
items/snippet/channelId parameter.
My current concern right now is that there is supposedly a way to get video information from the videos.list endpoint using a channelId, which does not seem the case in the API Explorer. I will update my answer once I figure out what this person meant exactly.
Looking at the YouTube API v3 Search documentation, it states that the order parameter is used to control ordering of API response resource results. However, when attempting to order by date, I'm not receiving the results in reverse chronological order as specified. In fact, they are not in any order whatsoever.
API Request with order date using API Explorer UI:
Is there some other way I should be requesting for results to be ordered by the time they were published on YouTube? Is there a bug in the handling of the order API parameter?
There's no bug; the 'order' parameter will sort results based on when the resources were created, This value may be slightly different than when a video is published, as a resource is created when an upload begins but a video is published when the upload finishes and is processed.
In this case, it looks like the 10 or so most recent videos were all uploaded at the exact same time (YouTube.com has a batch uploader), so you're getting the order they were created in, even though the published dates reflect minor differences (maybe a couple took a bit longer to upload or something). IF you scan the full list of results, though, you'll see they play out just fine in terms of their reverse chronological order. And generally you probably won't have so many videos that were batch uploaded like this.
I've explored most of this topic using the previous question How do I get a list of uploaded videos for a certain channel with the new YouTube Data API (V3)?, but I wanted some specifics that build upon the answer given.
Is it possible to use the channel's username without calling a search query to get the channel ID? I have an https request that can get a search list shown below, but I'd rather not waste API quotas when I know the exact channel username.
https://www.googleapis.com/youtube/v3/search?q={channel_username}&key={my_api_key}&part=snippet
My hack for this is just take a videoid from the channel and extract a uploader ID from there, but its not elegant and won't work if the upload is removed.
Also, once I've worked my way down to getting the uploads key, how can I get a json that shows all of the uploads and not just a list of 25? If I do have to recursively call pages of uploads, how can I do that?
Thanks.
I think that to get a channel, using the channel's username, the only option is to use the search, and the best strategy to minimize the use of quota is to use a cache, or save the channel ID.
For a full list of videos uploaded, as JSON, you can use contentDetails.relatedPlaylists.uploads (the ID of the playlist that contains the channel's uploaded videos) and retrieve the PlaylistItems, using pages with 50 items per page, and using the nextPageToken value to retrieve each page in the result set.
I've found the answer. The correct way to go about it is to query the site using the search, but directly state the type of list to respond with, in effect searching only channels. This is done by
https://www.googleapis.com/youtube/v3/search?q={search_term}&key={API_key}&type=channel&part=snippet
As you can see the type is channel. By searching with a search term of the exact channel username, you'll get a return list with the first item object as the channel, and you can store the channel ID and upload ID from there.
To update this answer, you need to recognise the "quota limits".
As such, to use the "search?q=[search]" will incur a 100 quota usage.
Better to use "channels?part=id&forUsername={username]" as is only a quota of 1.
Then you can use "channels?part=id,snippet ....... &id={channel_ID]" to grab all your data