30

I often search for closed questions and sift through the ones that are not duplicates to see if any of them need to be killed. Technically, closure is the first step towards deletion and the only closed posts that are actually useful are duplicates (unless it's a crappy one). For the rest, if the community hasn't bothered to reopen it after a while or if the OP has made no effort to edit their question, I delete it if it has no redeeming content. For example, the only closed posts seen on Gardening & Landscaping are exact dupes.

But this does become tedious when there are several results for various close reasons with duplicates interspersed between them. I'd like to be able to search as:

closed:1 duplicate:0

so that it gives a list of questions closed as OT/NaRQ/NC/TL. This will be similar in behaviour to migrated:0|1.

This will also be useful for 10k+ users on all sites (2k+ in beta) in the network, especially SO, so that they can search for questions to spend their delete votes on— especially closed, non-dupe questions in the tag they specialize in. More often than not, I hesitate to vote on questions that are outside my area of interest (unless it is crappy).

Improve this question
10
  • 4
    I just don't think this would be used enough to justify it, it's possible but we'd have to get a lot of feedback on how much it'd be used. We don't want to add an infinite number of operators, especially since search is getting a major overhaul underneath (it's my current project, and even that's just phase 1)...only what's absolutely needed will likely be added at the moment.
    – Nick Craver
    May 2, 2012 at 0:05
  • I suppose Nick would have mentioned it if it were relevant but I was under the impression that dupes were not treated specially -- just a blurb added to the post, rather than a field in the DB -- which would make such a search error-prone (needing to rely on the post content which can be altered or inserted into non-dupes). Perhaps I'm off-base though.
    – user154510
    May 2, 2012 at 0:18
  • 1
    Well if you could exclude terms that would help since they all have the same chunk
    – random
    May 2, 2012 at 0:25
  • 2
    @NickCraver Fair enough, but as random said, if you at least implemented the ability to exclude certain terms (which will have a much wider use than this narrow request), then this is a settled thing, because you could simply search for closed:1 -"possible duplicate"
    – Lorem Ipsum
    May 2, 2012 at 0:47
  • meta.stackexchange.com/questions/19253/…
    – random
    May 2, 2012 at 1:18
  • 1
    @MatthewRead - Yup, that's also true. It's not a field on the Posts table as all other related bool operators are...but due to how we index, it wouldn't be a tremendous burden to figure out...so I glossed over that tidbit.
    – Nick Craver
    May 2, 2012 at 2:09
  • 1
    @yoda - It's possible that search may work soon, I'll see what I can do. I'm in the middle of a major search API overhaul (we're moving it off the web tier to a service model) in preparation for [redacted]...if I can make that work, I will.
    – Nick Craver
    May 2, 2012 at 2:11
  • @NickCraver Thanks! Also, feel free to answer the question that random linked to with your comment, for the bounty :) (not that you need it)
    – Lorem Ipsum
    May 2, 2012 at 3:03
  • This isn't a solution mind you, but for now, this might help: stackapps.com/questions/3362/…
    – LessPop_MoreFizz
    May 2, 2012 at 3:48
  • @LessPop_MoreFizz Thanks, that will be handy
    – Lorem Ipsum
    May 2, 2012 at 4:26

1 Answer 1

Reset to default
10

Update: A change just went out enabling this on Stack Overflow and Meta:

duplicate:1 and dupliate:0 (true/false, yes/no also supported)

You'll notice the results start populating as the re-index propagates over the next 30-60 minutes. This requires an index rebuild which we will be performing after an elasticserch 0.90.5 upgrade on Thursday evening (around 7 PM EST), so all other sites will get it then.

There will also be a brief outage of search during that upgrade (around 5-10 minutes) due to the upgrade being cluster protocol shifting (requiring the entire cluster to go offline at once...or risk split brain). I'll post another notice of this on our @StackStatus twitter account and stackstaus.net closer to the outage.

Improve this answer
6
  • 12
    I think this no longer works with the new dupe message, right? I can get old closed dupes, but I don't think dupes with the new banner will show up. [status-unpleted]?
    – Ben Brocka
    Feb 8, 2013 at 19:08
  • This is definitely needed for beta sites which are not able to be searched on the data explorer or other fancy tools. Do we really need to run an API query to figure out if something is a duplicate? That seems like a doozy of a hurdle for growing communities to make the site a more accessible resource...
    – jmac
    Sep 9, 2013 at 1:01
  • 2
    Hi Nick, I've unaccepted and downvoted this answer. I suppose your workaround was kinda sorta useful then, but with the new duplicate banner, the text is no longer in the post, making this not useful. I think a dedicated duplicate: 0|1 operator as asked for in the question might be a better solution.
    – Lorem Ipsum
    Sep 12, 2013 at 19:48
  • 2
    @yoda Currently this would be crazy expensive, but I'm looking into schema changes that would make it and other things cheaper at the same time for core, mobile, and search.
    – Nick Craver
    Sep 12, 2013 at 20:18
  • @NickCraver Alright, thanks for the update.
    – Lorem Ipsum
    Sep 12, 2013 at 20:36
  • Aah, excellent!
    – Lorem Ipsum
    Sep 17, 2013 at 23:38

You must log in to answer this question.

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