pablobarbera / Rfacebook

Dev version of Rfacebook package: Access to Facebook API via R
http://cran.r-project.org/web/packages/Rfacebook
350 stars 250 forks source link

getPage fails to get all posts (skips some periods) #113

Closed nlch closed 7 years ago

nlch commented 7 years ago

I've been using the package to download posts from public newspaper pages and I've run into a small issue (other than that it's been working wonderfully).

It seems that when I try to download posts within a certain period, there are significant gaps in the download (for example if I download from june 2016 to septmeber 2016 I get all posts on june but it skips july completely and half of august). If I check the facebook page on that missing period there are many public posts that are not being downloaded.

Small example:

284515247529 should be Corriere della Sera

posts <- getPage(page = "284515247529", token, n = 1000, since = "2016-06-01", until = "2016-07-30", feed = TRUE, reactions = FALSE)

The response download posts from June but skips July.

kevin0932 commented 7 years ago

Hi @nlch , did you figure out why the function getPage() skips some periods? I met with similar situation just starting from 5 days ago. I am really appreciated if you can share some information or solutions you have got.

when I now use the function getPage() with no time filter, I can only get several hundreds of posts (used to be thousands, much more than those I could retrieve now) and they are in two clearly separated time ranges, for example, some old posts created before 10/12/2009 and the rests are between 10/04/2017-today 01/05/2017.

Thank you so much.

nlch commented 7 years ago

Hi @kevin0932, I haven't been able to figure out what the issue is. At the time I tested different things and I still had missing periods (which I only noticed because I had downloaded the same data before). I can't be certain but I think this happened right after one of the Graph API updates, yet I didn't find any mention of this anywhere else. The strange thing is, that when I tried to get the likes and comments of posts in those blank periods by their IDS I could gather that information with no problem.

I would have to try again and see what happens now without the since-until parameters tho... Hopefully that's not a permanent change.

kevin0932 commented 7 years ago

Hi @nlch , rodik pointed me to a post and seems that it is a bug after updates. https://developers.facebook.com/bugs/389409684792879 Seems that the bug has been fixed :) But I haven't tried it yet.

nlch commented 7 years ago

All the tests I did point to the bug being fixed :) Hopefully I didnt miss anything