Sometimes two feeds point to the same article. This can be irritating to display when the same article gets displayed several times. Can these be combined somehow during display?

Comments

alex_b’s picture

Title: Merge Duplicate Articles » Cross feed deduping
Status: Active » Postponed

There is no way to combine them. This feature isn't straightforward to implement (I've done it for feedapi_mapper and feedapi_fast) and there are no plans to do so at the moment for Feeds module in Managing News.

Jazz88’s picture

Pity, cause I think it would be a useful feature (maybe in the future). I can imagine that's not easy since many different factors would be involved.

Is there anything in the interim that could be done? Maybe just checking the referring url's??

alex_b’s picture

#2: You could extend Feeds module to support cross feed deduping (start with looking at FeedsDataProcessor class).

Jazz88’s picture

thanks alex_b, I will take a peak once i understand the feeds module a bit more...

greggles’s picture

I just reviewed the settings at admin/build/feeds/edit/syndication/mapping and noticed that it has "Item GUID" as a "unique target." Shouldn't that ensure that the data is unique across multiple feeds? Assuming that the GUIDs are properly created, of course....

alex_b’s picture

#5: it doesn't. For simplicity's sake it checks for an existing item for the current feed nid (FeedsDataProcessor::existingItemId()). The problem we have to solve if we do check for *any* existing item is that if we find one we need to associate it with the feed we're checking for. Otherwise feeds won't show items that actually exist on them but have already been imported through another feed.

tdombos’s picture

Yes this is a problem, but to tell you the truth the feed source is not a very interesting information in most cases, especially if you are following aggregated feeds. It does not really matter if it comes form feed A or feed B, it is the publishing site that matters. This is why I found Tattler (App) to be a more usable software (but not very mature in its internal workings, MN seems more well thought through) as they have the automatic story grouping by source (based on the domain the story comes from: you can map several domains to one source). Maybe admins should be given the possibility to choose between deduped stories and grouping by source URLs or no deduping and grouping by feeds.