0

On the Internet Archive data dumps for Stack Exchange, there is a file called PostLinks.xml. I know from viewing the files it contains the following fields: Id, CreationDate, PostId, RelatedPostId, and PostLinkTypeId. PostLinkTypeId can have values of either "1" (signifying Linked) or "3" (signifying Duplicate).

I am not clear as to what the purpose of this file is. It seems it allows one to identify linked or duplicate posts. Why would a person want to identify such things?

I also noticed that closed forums (i.e. closed private betas and closed public betas) do not have the PostLinks.xml file in their data dump. Is there a reason for this?

Improve this question
5
  • 1
    Identifying duplicate links is very valuable if you are analysing the data set. Tracking linked posts can be valuable too if you are analysing post relationships. I'm not sure what kind of answer you were expecting there.
    – Martijn Pieters
    Commented Dec 4, 2014 at 12:25
  • As to why closed site datadumps are missing those files; if that is true for all betas that'd be a bug, most likely. The possibility exists however that the site was closed before any duplicate closures had taken place.
    – Martijn Pieters
    Commented Dec 4, 2014 at 12:27
  • Thanks! Perhaps I should clarify: what is meant by a "duplicate link" and "linked post"? I am assuming a duplicate link indicates a question has been answered before while a linked post indicates a responder believes another post has relevance to the answer provided?
    – gajooge
    Commented Dec 4, 2014 at 12:29
  • 1
    A duplicate link is created when a post has been closed as a duplicate of another. A related post is created when another post is mentioned (in the question, the answers on the question or in the comments).
    – Martijn Pieters
    Commented Dec 4, 2014 at 12:31
  • 2
    Also see Database schema documentation for the public data dump and SEDE (and watch the sidebar on this post to see the relationship show up; the Linked section there are the related posts).
    – Martijn Pieters
    Commented Dec 4, 2014 at 12:32

1 Answer 1

Reset to default
0

Why would a person want to identify such things?

For example, if you wanted to import the SE data into your own website in some form, it would help to know which posts are duplicates of each other, and which are related in some tangential way. Having these links improves the structure of the knowledge base.

It is true that most people have no use for this file; indeed, most people have no use for the SE data dump at all.

The difference between two types of links was already explained by Martijn Pieters

A duplicate link is created when a post has been closed as a duplicate of another. A related post is created when another post is mentioned (in the question, the answers on the question or in the comments).

Improve this answer

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .