Thank you very much for your interest on the importer. I will try to answer your questions.
> 1) I read that the imports are for Likers and Commenters on the Page posts and nor for the page itself - correct?
Yes that is correct. Facebook doesn't allow anyone to download the full list of people who like a page or the members in a group. So these are the only data we can download.
> 2) One of posts on the Page I want to import happened on 23rd March (800+ Likes) and another on 1st April (500+ Likes). These are the 2 most recent posts. I import data for the page with Network set to "Co-Likers Network" and Options set to "Download
1 most recent posts" and both Get Status Updates and Get Wall Posts - UNchecked. It imports some data with 300+ vertices. What is data for? As I said the 2 most recent posts are on the given dates..
As I can see you downloaded only 1 post and the most recent one which is the post on the 1st April. The reason why you got 300 vertices and not 500 is because of the privacy settings of different users. Since they are not your friends (you being the logged
in user) and they have set their settings to not be discoverable by other people or other people can not get extended information for them, Facebook doesn't give us their information. Also I can mention here that sometimes even if you are able to see some
very basic information for a specific user (i.e just the name and sex) on Facebook, you can not access the same information through the API. However the same thing may happen in the Co-Commenters case and in this case there is another explanation. What you
actually see on a post on Facebook is the number of comments and not the number of commenters. This may be lower since a person can make more than one comment on a post).
> 3) I changed the Options to "Download posts from 23rd March 2012" with all other filters as above. I was expecting this to get me at least the 800 vertices from the 23rd March post. But this query never goes through and I get a Java OutOfMemory exception.
I have imported huge Twitter networks with more then 30,000 vertices in NodeXL itself. So I cant believe the problem is to do with the quantity of data am accessing through this query. Is it a bug or am I doing something wrong?
Actually that is a problem of NodeXL. The Co-Liker network is a little bit tricky since it will construct a clique (complete graph) for all the likers of a post. So for the post on 23rd of March you have 800 vertices (which is a small number), but it will
produce a network with 800*799/2=319600 edges. And that is for a single post! Also the Co-Commenter network is tricky because it has the same logic, but usually there are less comments than likes and most of the time we get more information
by analyzing commenters rather than likers.