Hi,

Pls give your thoughts/suggestions to backup chatter data. Following is my current strategy.

Of the various aggregate feed objects like NewsFeed, UserFeed etc UserProfileFeed is the one most useful for Chatter backup. UserProfileFeed gives us the entire information about a users activity related to chatter. This is given by NewsFeed too but NewsFeed additionally gives us information about the users/objects followed by a particular user. If we use NewsFeed for Chatter backup then we will have lots of duplication.

The following query gives us the entire chatter data for a particular user.

Select Id, ParentId, FeedPostId, Type, CreatedById, CreatedDate, IsDeleted,
LastModifiedDate, SystemModstamp, CommentCount, LikeCount, Title, Body,
LinkUrl, ContentData, ContentFileName,ContentDescription, ContentType,
ContentSize, InsertedById , (SELECT Id, FieldName, OldValue, NewValue
 FROM FeedTrackedChanges ORDER BY Id DESC), (SELECT Id, CommentBody,

CreatedDate, CreatedBy.FirstName, CreatedBy.LastName  FROM FeedComments

ORDER BY CreatedDate),  (SELECT CreatedBy.FirstName, CreatedBy.LastName 

FROM FeedLikes) FROM UserProfileFeed with userId='<user-id>'

The above query can be optimized by observing that the records for which feedItem.Type='TrackedChange' can be removed if they have no comments made on them i.e FeedItem.commentCount > 0 then and only then include them for backup.

Additionally  we will need to back up the EntitySubscription object so as to capture
all the follower-followee combinations.

And finally these collab objects need to be backed up too. CollaborationGroup and CollaborationGroupMember.

However one   question I am struggling with is this.
What if the parentId points to an object instead of a user?

Then should the Chatter backup also backup the entire object too.