Podcast #128: We chat with Kent C Dodds about why he loves React and discuss what life was like in the dark days before Git. Listen now.

Hot answers tagged

9

Inline tag edit requires 10k rep, but 2k rep is enough to edit questions (therefore, do retags). However, inline API is much easier - you just do the post, which contains tags only. Doing full edit requires accessing postable questions body. I've done quick modification to your script. It's a draft, but working for me... it activates edit mode, waits until ...


5

the original question was apparently for V1. in V2 we now have https://api.stackexchange.com/docs/required-tags the request: https://api.stackexchange.com/2.0/tags/required?order=desc&sort=popular&site=stackapps will return required tags


3

Yes. I would agree that the tag serves no purpose. With only 24 questions, let us lose it.


3

This is a bug and seems (based on a series of A/B tests) to be linked with the body_markdown property. That is, if your filter returns the tags array, it works: "items": [ { "tags": [ "batch-file", "cmd" ], "answer_id": 42292030 }, { "tags": [ "ios", "arrays", "swift" ], "answer_id": 42295110 }, etc... , unless body_markdown is also ...


3

This is very costly to do in the API and currently impossible for many combinations of tag and date (because it would require more than the 10,000 allowed API calls per day). I'll write more about that tomorrow (11/11 or 11/12 UTC). Meanwhile, perhaps a SEDE query will work? The query: Top Answerers, by your tags and your date range takes your date range ...


3

I'm new to the API also, I think mostly multiple args are separated with ; Try c;java


3

Related tags were one of the "maybes" slated for API V2.0. It got implemented today. Note that our definition of related is pretty much "appears on the same questions as", there's no real secret sauce there.


3

Based on a reasonable response for Stack Apps, this has now been done. Merged stack-exchange into support (which many of the questions already had) and deleted support from those posts where it didn't apply. These 24 questions were affected:     2332, 2371, 2372, 2373, 3045, 3136, 3408, 4457, 4617, 4699, 4888, 4942,     6155, 6189, ...


2

At the moment the number of tags allowed on a question is 5. Both client side and server side this limit is now1 enforced at the time of posting. There is no direct limitation to allow 6 or even 7 tags. As a matter of fact, this query shows 1038 questions at the time of posting that have 6 tags. The revision for one occurrence reveals what caused that 6th ...


2

As a result of some typos, those fields were being queried for... and then discarded immediately. A fix has been deployed.


2

It's because tag ids aren't stable, so you can't reliably use them in the API. The same reason sites don't have ids. You'll note that most everywhere in the site's when we refer to a tag, we refer to it by name. Only when editing a few things do we expose the tag id, and those links can safely break.


2

The tag object now has a has_synonyms field which is true when there's at least 1 approved synonym for a tag.


2

No, we won't be implementing this. Not that I don't see a use case for it, but we'd never be able to accept and execute arbitrary regular expressions against the database. It'd just be far too easy cripple performance wise (we've done this accidentally plenty of times, and we don't have to worry about malicious actors on the dev team). While we could ...


2

No, there's no API route that gives you a user's tag specific reputation information. The /users/{ids}/tags route only returns the combined count of questions and answers that a user has posted in a tag, but not the reputation information that you can get on the user profile tags tab. (Note: There's an app at Badge Status for a particular Tag in ...


2

Hear! Hear! I want to study what makes a good answer on SO. I started by querying complete web pages (questions with all comments, answers, and their comments) for questions with some minimum score (such as 5) for a number of fixed tags A,B,C created in some time period (say, one full week two months ago). This can be done well with \search and a filter ...


2

First, you'll want to go upvote the feature request to add tag search for answers, etc. (If/when you have the upvote privilege) In the meantime, for answers by tag you can use the /search/excerpts route as long as you include is:answer in the q parameter. For example:       /2.2/search/excerpts?&q=is:answer&tagged=best-in-place&...


2

The question was not originally clear if you wanted tags or badges, but the answer is the same... To just get totals, use the total filter. For example, for badges: /badges?site=stackoverflow&filter=total, which currently returns: {"total":6363} For tags: /tags?site=stackoverflow&filter=total, which currently returns: {"total":48667} WARNING: ...


2

Do you mean something like this plot? : (Generated by SEDE: /query/201360?ShowScore=0&Tag1=excel&Tag2=python&Tag3=NA&Tag4=NA#graph) Use the data explorer (SEDE) for that. To do this with the API, is possible but not pretty. For example, to replicate the above plot you would need to make 111(Age of SO in months) API calls for python like ...


2

No, the API cannot do this. Refer to the API docs and you'll note that there is no method that can be used to compute a user's tag percentile‡. Also, it's not clear exactly how it's calculated: Many tags are not eligible. (Number of questions? Total tag score?) It might be that users with less than 200 rep are not counted. Anyway, you can get the ...


1

No, this information is not exposed via the API (and it's not available in SEDE either). The same question was posted here on Meta Stack Exchange. User @rene found a way to scrape it from the website; see here for his solution. It could be that the rename / new functionality (tag favorites -> tag watching) broke it, I haven't tested it.


1

From the /questions doc: ...use the tagged parameter with a semi-colon delimited list of tags. This is an and contraint(sic), passing tagged=c;java will return only those questions with both tags... (Emphasis added) If you want questions that have either tag, use the /search route (or /search/advanced). For example: /2.2/search?tagged=google-docs-api;...


1

There is no need for a special API route, "OR" logic is easy to implement. Merely search for each term separately and then merge the results. For example, if I wanted tags that had either javascript or for-off, I'd first search for javascript(44 tags), then for for-off(2 tags), and then merge the results to get the 45 unique tag names. Here's a ...


1

You can do this on the API with far less than 120K requests. Just use a different approach. Instead of getting a total for every single month, offload that work to your app. Use a /search query to get the questions for the whole interval in question. EG: /search?pagesize=100&fromdate=1329523200&todate=1487376000&sort=creation&tagged=...


1

There seem to be 2 or 3 effects, the biggest is that the Tags page also shows Synonyms. There may also be a small cache effect, but there seems to be another discrepancy as well, possibly an error/bug. You can see that the Tags page also shows Synonyms by: Go to the Tags page. Enter userscr in the Type to find tags: box. You will see something like: Then ...


1

The API is not a good tool for this (more below). Use the Stack Exchange Data Explorer (SEDE) for this kind of bulk analysis. For example, this SEDE query shows the number of unique users who asked and/or answered in the PHP tag: Unique, Overall, Active Users: 384,835 Active Question Askers: 256,140 Active Answerers: 209,108 ...


1

While this isn't possible with the API, it is possible with SEDE, if you allow for up to 7 days of lag. Here is the SEDE Query that lists Tag wiki and excerpts in descending edit date order: select t.id , t.tagname , concat('site://tags/' , tagname , '/info' , '|' , case when w.lasteditdate > e.lasteditdate ...


1

To do this you have to use the /users/{ids}/answers path to download all of a user's answers, and then filter out the tags and/or tot-up the scores from there. Pseudo code: scoreByTag = [] // Empty associative array while (more pages of results) { 1) Fetch next page of answers 2) Store results and/or update scoreByTag array } Display/use select ...


1

The search path already allows more than 5 tags, and it uses OR logic. From the docs: tagged and nottagged are semi-colon delimited list of tags. At least 1 tag in tagged will be on each returned question if it is passed, making it the OR equivalent of the AND version of tagged on /questions. (bold added) Note that tags are separated with a semicolon, ...


1

For questions tagged with all tags in a list (AND operation), use the /questions path. For example: /questions#fromdate=2013-01-11&sort=activity&tagged=ruby;git ~~~~~ For questions tagged with any tags in a list (OR operation), use the /search path. For example: /search#fromdate=2013-01-11&sort=activity&tagged=ruby;git


1

Kevin was nice enough to grant my recent feature request on this endpoint, so it now returns "paged" results instead of a static top 30. I'm guessing the documentation just wasn't updated to reflect the change. I've tested it (StinkingBadges is using this functionality) and can confirm that your test result is the current "correct" behavior.


Only top voted, non community-wiki answers of a minimum length are eligible