6

My apologies if this question is off-topic here, but it is a discussion about how to use a network-wide tool specific to SE.

The question in its entirety is in the title, really. How can I use the StackExchange Data Explorer to run a query across all sites in the network? I cannot seem to find a way to search more than one (set with the drop-down next to Compose Query) – is this or is there a workaround?

Improve this question
0

1 Answer 1

Reset to default
6

Yes! You just can't use it, because it's not enabled on the public-facing instance. This is indeed .

Technically speaking, you can hack your way around this*, but I don't recommend it. It's pretty rare that there's a good non-community-management reason to run a query against every single site at once, so being able to switch sites relatively conveniently should be sufficient for most cases. I'm opening to hearing counterexamples, although I don't control what features are enabled on the hosted SEDE.

*though not exactly as shown there, because EmailHash no longer exists, and that was for joining cross-site, which is more than what you were asking for anyway

Improve this answer
6
  • For vermiculus/sx.el#137, I would like to get a rough estimate on how many tags exist throughout the network to estimate how many API calls it would take to retrieve them all. Since there are just too many to grab for each user (that would be irresponsible), we need a workaround if we want sane tag completion. The idea is to create a bot that goes through and grabs tags by the batch and dumps them to a file that can just be downloaded and read by the client, but this still relies on thousands of API calls.
    – Sean Allred
    Dec 5, 2014 at 2:09
  • What may have been a better solution if this were not status-bydesign would be to run the query once a day or so and scrape the output. If SEDE runs off of the quarterly dump, though – as I seem to remember it does (?) – I don't know how much that's going to do.
    – Sean Allred
    Dec 5, 2014 at 2:12
  • SEDE is updated via a separate process that runs weekly, though that still might be a too-large window for your purposes. For your estimation purposes, you could write a quick script to sum up tag counts on the SEDE homepage. It does seem like the API should have exposed a tag searching endpoint though, hrm...
    – Tim Stone
    Dec 5, 2014 at 2:26
  • Oh goodness, that was our fallback, though: making a request to the API to search for tags based on current input. I guess that's out :(
    – Sean Allred
    Dec 5, 2014 at 2:28
  • 1
    Ah, I misspoke on that one, the /tags endpoint does allow searching via the inname parameter, so that seems like your most viable option.
    – Tim Stone
    Dec 5, 2014 at 2:33
  • These are in the space of "just curious" rather than "urgent" or "would majorly improve my ability to contribute to SE", but two queries I'd like to be able to run are: (a) on which sites does $user have $badge, and (b) what are the top N sites for $site's top M users (i.e. where else do these users participate significantly).
    – Monica Cellio
    Dec 5, 2014 at 3:24

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