This is an archive of past discussions about Help:Searching. Do not edit the contents of this page. If you wish to start a new discussion or revive an old one, please do so on the current talk page.
I think I'd better say clearly that recent changes I made actually change what users are recommended to do. The document essentially said "spaces are optional in most cases, but in a few instances must not be present" (e.g., in, Insource:). I think that is confusing. There is no reason to type spaces if they're not needed, so I've changed this to say essentially "don't use spaces if they're not needed, but they're usually (not always) ignored if you do". Full details were, and are continued to be, supplied; it's just the recommended use that's changed. I've also removed some unnecessary spaces from examples. I've already done this, and nobody seems to be objecting, but I thought I'd better point it out. Best wishes, Pol098 (talk) 19:00, 16 December 2016 (UTC)
Edit Request
This edit request has been answered. Set the |answered= parameter to no to reactivate your request.
"WP:S" redirects here. For the Sandbox, see WP:SB.
The first hits are for things like the article on quotation marks. This occurs regardless of the contents of the strings inside the quotation marks. Ideas?Naraht (talk) 16:06, 6 January 2017 (UTC)
Considering adding the following at the end of section 'From any web page'
In you use the Chrome browser with Windows then you can add a setting to the browser's context menu that enables you to use google to make a restricted search of wikipedia from any webpage for words that you highlight in the page.
From any page, highlight a word (using either mouse or keyboard gestures).
Now, either right-click or Shift+F10.
Select menu item 'Context Menu Search' with either mouse-click or arrow keys then Enter.
Select menu item 'Options', at the bottom.
See a form entitled 'Add more Search Engines manually'
In the 'Display name' field enter en.wikipedia (or a name for this item of your choice).
Now, whenever you want to search for a word or series of words in any page from within Chrome simply highlight the search item, right-click on it, select 'Context Menu Search' and select 'en.wikipedia'.
My questions
Any objections?
A better place for the inclusion of this, assuming it's unavailable elsewhere? — Preceding unsigned comment added by SewerCat (talk • contribs)
@PrimeHunter: I appreciate your thoughts! I was aware of the fact that Google can be an option in the search box but I find this much easier to use and I thought that others might too. Missed the Google Canada entirely. Can you given me some general rule for reducing the level of detail. I agree about where this should go. SewerCat (talk) 14:25, 14 May 2017 (UTC)
New search results feature
In the ongoing effort to improve the discoverability of all the world's knowledge, the Discovery team is proposing an enhancement to the search results page on Wikipedia (Special:Search). The goal of this feature is to display related content for each individual search result returned from query, so that even when the search result itself isn't maybe entirely useful, perhaps its related content would be. We would like your early feedback on this new feature—more information can be found on MediaWiki about the explore similar functionality and testing can be done in your own browser, using step by step self-guided testing instructions. DTankersley (WMF) (talk) 21:00, 31 May 2017 (UTC)
See [1]. Basically, if a page has a multi-word title, and you search for an exact one of the words in that title, you'd expect to find that page in the search results, but it is rarely even on the first results page. Instead you get lots of irrelevant stuff that isn't spelled the same way. Is this a bug? What can we do? Equinox◑21:23, 12 July 2017 (UTC)
Hi, Sorry for sounding stupid but how does one search for the most recent results ?, -recent and prefer-recent doesn't work for me? (IE when I search ""London" -recent" or ""London" prefer-recent" I still get jumbled results), Thanks, –Davey2010Talk21:28, 12 June 2017 (UTC)
Question about incategory (or category related) search
This is after I asked this question and got an answer here. However I either didn't ask the question correctly, or I simply haven't been able to use the answer to get to where I need. I want to search for all articles (not categories) that have been assigned: 1. any category whose title includes the term "Scorsese", boolean-AND 2. any category whose title includes the term "comedy". I tried advanced search by checking only the box for Category but this is for category results, and there is no single category that includes both these terms. I checked also the Article box, but I got many unwanted results. Then I tried various combinations with filters and boolean operators but I can't seem to get what I need. Any further clues? Thank you. HoverfishTalk22:36, 18 July 2017 (UTC)
Thank you very much Izno. I also use PetScan and I see what you mean with insource + regex, I hadn't thought about it but I will try it next. My main concern however is not so much for editors tasks but for a way to make such searches more approachable to the layman. I will check RAQ and see if I can find some way to encapsulate such querries. For WP Films it's only a certain number of main-film-genre-terms that have been diffused in too many subcategories to be usefull in searches. Thanks again for the tips. HoverfishTalk14:00, 19 July 2017 (UTC)
Semi-protected edit request on 8 September 2017
This edit request to Help:Searching has been answered. Set the |answered= parameter to no to reactivate your request.
"Sister projects" should be changed to "similar projects"
...or something that is likewise non-genderized. This turns up to the right on the search page where the suggestions are offered. It sounds off and awkward and inappropriate. Like something from the ministry of feminist information, like some kind of female only endeavour. Please stop this double standard. If you have to say "humanity" or "humankind" instead of "man" or "men" then it's only fair to stop this "sister" nonsense too. Gender bias free language should apply all around. Thanks. Alialiac (talk) 17:32, 7 October 2017 (UTC)
Searching within (or exclusively without) reference contents...
Is there any way to limit a search to specifically text within a <ref></ref> pair or specifically to everything that isn't inside them? I know regex, and I'm sure that a sufficiently complicated insource regex could do it, but I'd rather not, if there was any other way.Naraht (talk) 15:49, 3 November 2017 (UTC)
Insource is the only way besides resorting to a database scan... which would probably also require regex. --Izno (talk) 17:15, 3 November 2017 (UTC)
That said, if you know the text internal to the reference, you can also search for that instead. (Unless you're looking for template parameters or something?) --Izno (talk) 17:16, 3 November 2017 (UTC)
Semi-protected edit request on 6 January 2018
This edit request to Help:Searching has been answered. Set the |answered= parameter to no to reactivate your request.
In case anyone has been wondering about the &searchToken=something at the end of the URL for a search result, I asked about it at mw:Help talk:CirrusSearch, and the summary is:
When publishing or saving a search link, there is no benefit in including the &searchToken=something at the end of the URL, so it's better left off.
The page has three links to {{search link}}. It's just a template at the English Wikipedia and not a part of the search feature. I don't think Help:Searching should explain how the template works. It's documented on the template page. PrimeHunter (talk) 22:48, 13 March 2018 (UTC)
Confusion
@Cpiral: I am confused by this edit. No, WP does not require Javascript. However, the placement of the emdashes indicates that you believe the "default" is Vector without Javascript. This is not the case for the majority of people. --Izno (talk) 13:35, 17 April 2018 (UTC)
You're right. The mdash phrase was a compact way of referring to multiple kinds of defaults: default behavior (when JavaScript is off) and default skin (Vector). I will improve it. — Cpiral§Cpiral17:58, 17 April 2018 (UTC)
Should Qwika be removed?
As I look at the Qwika search page, it is showing "Updated Crime Data!" with a date of "Thu 31 Mar 11", which I take to mean that their crime data information was last updated 7.5 years ago. If the rest of the search engine is that old, perhaps the Qwika section should be removed from the Help:Searching page. Eddie Blick (talk) 01:05, 28 August 2018 (UTC)
Thank you, Izno. Yes, it was the External search engines page. I had both pages open and typed the wrong title in my comment. Thanks for catching my error. Eddie Blick (talk) 01:54, 28 August 2018 (UTC)
Semi-protected edit request on 11 December 2018
This edit request to Help:Searching has been answered. Set the |answered= parameter to no to reactivate your request.
Main revamp is complete. I'll be back to work on this page some more. Though feel free to jump in and improve the page. The more the merrier. Below is a task list of work needed on the page. Feel free to add to it. — The Transhumanist08:16, 16 March 2018 (UTC)
Place content from overview section Done
Identify the search engine Done covered in the section "Under the hood"
Integrate merge section under search box Done
Integrate merge section under search results page Done
The lead needs refining, and section linkification.
Namespace and sister project prefixes should be covered, or at least a link to the help pages should be provided.
A full proofread is needed.
Check these facts (include them if they verify as accurate)
Each parameter has search indexing and page ranking prerequisites
A regular expression search is not an indexed search, and so in order to pre-establish its reasonable search domain, a regular expression search requires an accompanying search term that is an indexed-search filter
It also offers one-button modification of the search domain, and through user preferences, the Wikipedia Search results page can run the search on other search engines, and go to their search results page.
When presenting results, the internal search understands and will link to relevant sections of a page
The ordering of the list of search results is determined by the page ranking software, but you can make adjustments to it. [How?]
For articles, a message box may show up beside a listing, indicating that it has a sister page on another project, such as a Wiktionary entry that defines the title of a Wikipedia article.
Track down the bug in regular search box pertaining to subpageof: parameter
How do I search for article titles containing the word "Strasse"? I tried intitle:Strasse and intitle:"Strasse" and intitle:/Strasse/; they all returned titles with "Straße". -- Michael Bednarek (talk) 01:24, 29 January 2018 (UTC)
Has anyone already done a JS script that adds some way to get directly to Advanced Search without first having to do a search? I almost never use the search box to just search for something, unless I"m in "reader mode" and am not working as an editor. My usual routine is to enter something like asdfjasodfijasdofj in the search box and search on it, then get into advanced mode, then pick my namespaces, then replace the search string with what I'm actually looking for. Rather inefficient. Hoping there's a "widget" of some kind that puts advance search as a clickable icon or something next to the default search box, in the site-wide top menu. — SMcCandlish☏¢ 😼 21:48, 13 July 2018 (UTC)
$.when(mw.loader.using(['mediawiki.util']),$.ready).done(function(){mw.util.addPortletLink('p-personal',mw.util.getUrl('Special:Search')+'?profile=advanced','Search','pt-Search','Go to advanced search',null,'#pt-logout');});
How to ask "is infobox parameter coordinates" controlled by wikidata values?
I asked this question (more or less) over at the WP:Helpdesk, where is has been greeted with '???!?!?' so far. It is an unusual search request!
Tsing Yi North Bridge used to show geophysical coordinates displayed twice. For reference, Tsing Yi South Bridge still does. The coordinates were coming from both an embedded
So the question is: how could I search for articles where the {{Coord}} template is used and where now Wikidata is enabled for the articles and has coordinates values? I can't see the ability to ask questions about Wikidata and a particular article.
The page "Universal welfare" does not exist. You can ask for it to be created, but consider checking the search results below to see whether the topic is already covered.
I want to find all instances of "Phi Alpha" on Wikipedia that are not part of "Alpha Phi Alpha" or "Phi Alpha Theta". But I can't get the look ahead or look behind syntax to work. Does Cirrus Search do look ahead/look behind examples? I know to pair whatever I can do in insource with simply searching for "Phi Alpha" in order to have it finish at all.Naraht (talk) 19:32, 8 July 2019 (UTC)
Please review proposed updates to boolean/logical operator documentation
As discussed in more detail in Phabricator ticket T228925, it turns out that parentheses and the boolean operators AND and OR do not actually work as advertised in the documentation. The short term plan is to update the documentation to reflect current reality, and the long term plan is to fix reality to reflect the currently aspirational documentation. Here's the first draft of the update to go in the Help:Searching Logical operators section:
The search engine supports limited boolean logic in searches. Logical NOT (negation) can be indicated by a "-" (minus sign) or a "!" (exclamation point) character prefixed to a search term, or by the NOT keyword.
Parentheses (…) are ignored by the search engine and have no effect.
The operators AND and OR are used by the search engine, but do not have the expected boolean logical meaning and should be used with great care. See the additional documentation for an in-depth explanation.
I've also written a draft of the longer explanation (~1200 words) of the use of Logical operators in on-wiki search. Feedback is much appreciated! I still need to review the rest of the Help:Searching page and mw:Help:CirrusSearch page to see if any other changes are needed there. TJones (WMF) (talk) 15:29, 30 July 2019 (UTC)
^ and $ in intitle regular expressions
I would expect intitle:/^Europe/ to show pages with titles starting with "Europe", but instead there are no results. Is the lack of support for ^ and $ in regular expressions a known limitation? Is there a workaround? falsifian (talk) 17:50, 18 August 2019 (UTC)
Thanks. I don't see support for suffixes (suffix:Europe seems to just search for "suffix") or for regular expressions in a prefix (prefix:/Europe/ returns no results), so I am still interested in a general answer. I might file a bug on Phabricator. For some context: a friend and I are planning to build a tool to support move request discussions (especially ones invoking WP:CONSISTENCY) by showing what kinds of title are in common use, or to find more titles for a multi-rm. For example, intitle:/Award$/ vs. intitle:/Awards$/ would have been relevant to this discussion. falsifian (talk) 18:10, 18 August 2019 (UTC)
I think the answer regarding suffixes (and I'd guess prefixes) is phab:T90090, though I think that comment was made before regex was implemented for title searching. There may also be some related chatter in phab:T12808. --Izno (talk) 18:20, 18 August 2019 (UTC)
The deepcat example (deepcat:"musicals") seems to be off. For some reason, if the category's first letter is not capitalized, the search doesn't seem to search in the subcategories. First see Category:Musicals and then compare these searches: deepcat:"musicals" vs. deepcat:"Musicals". The latter search also gives the following warning: "A warning has occurred while searching: Deep category query returned too many categories". Maybe we should change the example search and category to a smaller category with correct capitalization, for example to deepcat:"Musicals by topic"? 85.76.139.129 (talk) 19:58, 12 July 2019 (UTC)
It says above, that this is not the place to ask questions. Where is it, then, that we ask questions about how "search" works, or more specifically, how it doesn't work? I have some queries that are not getting the desired and expected results. Thanks. Gah4 (talk) 02:35, 17 October 2019 (UTC)
OK, I figured out the answer without so much help from this page. I am searching for something with a # in it. It seems that I need insource:/.../ to get a regular expression, and \# inside the regular expression to match the #. — Preceding unsigned comment added by Gah4 (talk • contribs) 02:42, 17 October 2019 (UTC)
Yup. You should prefer to search for some other text though as well, since regex insource searches are expensive. --Izno (talk) 04:18, 17 October 2019 (UTC)
It seems that it wouldn't match the # without it. And I didn't do it so many times. If it should match the #, then maybe it needs fixing. Gah4 (talk) 08:28, 17 October 2019 (UTC)
It's common for search features to ignore punctuation characters, also in quoted strings. Google does this. Izno meant a double search like at mw:Help:CirrusSearch#Regular expression searches where another part of the search reduces the number of possible pages so the regex part only has to be applied to those pages. PrimeHunter (talk) 10:40, 17 October 2019 (UTC)
It is common, but the documentation says that inside quotes that they should match. The actual search is: insource:/List_of_Nobel_laureates_by_university_affiliation\#/ so no metacharacters, and I believe fast enough. The similar search with double quotes instead of / doesn't work. Thanks. Gah4 (talk) 19:41, 17 October 2019 (UTC)
After I did the above search, I did the same with blanks instead of underscores. I hope there aren't ones with mixed blanks and underscores. With quotes instead of /, it ignores both # and \#. I haven't tried any double search like you show. Tradition of GNU grep is to do a Boyer-Moore search on the longest fixed string before doing the regexp match. I suspect others now do that, too. Is there a place, other than here, where we should ask these questions? And thanks for answering them here. Gah4 (talk) 01:36, 18 October 2019 (UTC)
Answering here is fine, though you could also ask at WP:Help desk or WP:VPT.
So, since you still don't seem sure of yourself: The # (like most punctuation) is not matched inside quotation marks because it is considered grayspace (at mw:Help:CirrusSearch#Words, phrases, and modifiers: []\:";'<>?,./. A mixed string of greyspace characters and whitespace characters, is "greyspace", and is treated as one big word boundary. Greyspace is how indexes are made and queries are interpreted.). See also Help:Searching#Search string syntax.
For caps insensitivity in a regex, you can do /i e.g. insource:/search_term/i as the slow route and [Aa] where desired for the faster route. You can't get the same for spaces versus underscores unfortunately, so you should use a group ( |_) or search on another term of interest such as affiliation, above. --Izno (talk) 03:17, 18 October 2019 (UTC)
Semi-protected edit request on 4 January 2020
This edit request to Help:Searching has been answered. Set the |answered= parameter to no to reactivate your request.
I think it would be really useful for the users. It can help them in many ways. For example to introduce the list in a text file and to further refine the search, using a text editor, or for using a robot for changing text in the respective articles (for example there are 237 articles containing the string "Neamţ County" and a robot should change the text into "Neamț County" - from T-cedilla to T-comma). Thank you. — Ark25 (talk) 21:44, 29 March 2020 (UTC)
It hides the other lines on search results pages. If you use Vector normally then you can change a search page to display in MonoBook by adding &useskin=monobook to the url. If you have MonoBook then place it in Special:MyPage/vector.css and add &useskin=vector to the url. You can also manually apply the above css to a viewed browser window with a feature in most browsers. You can also add this to your common JavaScript:
$(document).ready(function(){if(mw.config.get('wgCanonicalSpecialPageName')==='Search'){mw.util.addPortletLink('p-tb',location.href.replace(location.hash,'')+(location.search?'&':'?')+'useskin=monobook','MonoBook','t-monobook','Repeat the search in MonoBook');}});
@PrimeHunter: These scripts are really great! I forgot to mention that I am not using AutoWikiBrowser; many of the users who would benefit from this feature won't have the time and patience to use AWB.
Do you have any link to some documentation that explains how to manually apply the above css to a viewed browser window ? I'm just curious.
@Ark25: It depends on the browser. In Firefox I right-click somewhere on the page, click Inspect Element, click Edit CSS, and copy the css code to the top. PrimeHunter (talk) 00:42, 30 March 2020 (UTC)
Escape characters
Is it possible to escape certain characters in the searchbox? I would like to search for intite:/, Ohio/ but it will ignore the comma character (,) so is it possible to escape this character? Like writing "\," instead of "," . Thank you. — Ark25 (talk) 21:05, 30 March 2020 (UTC)
@Ark25: You are missing "l" in intitle so it makes a normal content search and not a regex title search. The correct intitle:/, Ohio/ does not ignore the comma. Backslash "\" is the escape character in regex searches but it's not needed for a comma. PrimeHunter (talk) 21:18, 30 March 2020 (UTC)
articletopic
The central help page now has a section about the articletopic: search keyword which has been added a few weeks ago. Maybe worth adding here as well. --Tgr (WMF) (talk) 19:28, 10 April 2020 (UTC)
Let's Eliminate or Document Limitations re. aliases and Pseudo-namespaces / pseudoaliases.
Search, and this page, are a nightmare when it comes to pseudo-namespaces. Let's Eliminate or Document Limitations re. aliases and Pseudo-namespaces / pseudoaliases!
In the H:PARAM Parameters section, it's written under
namespace name:
...Namespace aliases, like "WP" for "Wikipedia", are accepted.
I claim most users would prefer if a search for T:dubious or H:PARAM didn't fail / did bring up template:dubious and H:PARAM. You would prefer otherwise? Why? Why would anyone prefer they not work as suggested. And who else do you claim to speak for?--50.201.195.170 (talk) 19:10, 6 January 2021 (UTC)
Search knows nothing about the idea of pseudo-namespaces, and I assume that also applies to most readers. Help:Searching is mostly for readers while pseudo-namespaces are mostly for editors. Mentioning pseudo-namespaces in Help:Searching might cause more confusion than it solves. Wikipedia:Namespace#Pseudo-namespaces says: "Pseudo-namespaces are not in any way recognised by the wiki software; they are purely a community custom. Titles in pseudo-namespaces actually belong technically in the main (article) namespace and are treated as such by the software: they are case-sensitive and appear in search results restricted to the main namespace." Pseudo-namespaces are merely a custom to make a limited number of mainspace redirects to other namespaces on names starting with certain strings like CAT:, H:, MOS:, P:, T:. PrimeHunter (talk) 09:16, 4 January 2021 (UTC)
Yes, we don't want to confuse readers... and H:Searching, which readers certainly do use, is itself in a pseudo-namespace. (Do recognize it is mostly for them.) You're quoting from a page I linked to. Aware of all that. So, what's your suggested solution? You are posting in the "Solution options" section...
I claim most editors would prefer if a search for T:dubious or H:PARAM didn't fail / did bring up template:dubious and Help:PARAM, and most readers won't do such searches. Agree?
Creating a "Solution options" subsection from the start was odd, especially without signing the main section. I'm not going to create a "Comments" subsection to address your posts (and I clicked edit at the main heading). H:SEARCHING got 10 page views in 2020 and Help:Searching got 240,568.[3] Help is a real namespace. I'm not sure what you want by "adding support for Pseudo-namespaces" if isn't to actually make them namespace aliases like WP. That could be done with mw:Manual:$wgNamespaceAliases but it would require developer support. T: for Template would make things a little easier for me but I can live with typing a few more characters. Most people would probably prefer T for Talk if it was a real alias. PrimeHunter (talk) 22:05, 6 January 2021 (UTC)
Most people would probably prefer if _t worked where _talk did? Isn't that more in line? True, T: for Template: would make things a little easier for me too. What what do you think it (H:SEARCHING got 10 page views) shows? Besides, I asked about H:PARAM, but you searched for H:SEARCHING. I see over 1400 page views for it alone: https://pageviews.toolforge.org/?project=en.wikipedia.org&platform=all-access&agent=user&redirects=0&start=2020-01-01&end=2020-12-31&pages=H:PARAM%7CH:PARAMETER. AND THAT'S JUST FOR TWO OF THE TWENTY-SEVEN shortcuts. What would be more relevant would be a search of search history, or a poll, but really, I would think being in our own shoes would work. I did describe I want by "adding support for Pseudo-namespaces". See "hack", now bolded, above. While "to actually make them namespace aliases" (and have the existing content align with that) would be how many orders of magnitude more work, do you think? --50.201.195.170 (talk) 04:43, 8 January 2021 (UTC)
Safari search?
The section on Safari search doesn't seem to work any more. Is there a new way to search Wikipedia easily within the address bar of Safari? Cheers. Wikidea10:54, 29 July 2020 (UTC)
Are there any best practices or policies I can refer to on how Wikipedia makes content easier to find using search? I’m thinking particularly about how pages are organised and structured.
For example, are there any best practices on how articles are titled, when a page redirect or disambiguation page should be used, how categories should be structured etc.
I administer an instance of mediawiki, and the biggest criticism I hear is that nobody can find anything with search. Although we’re not using CirrusSearch, I largely put it down to a content structure that wasn’t well thought out and is governed poorly. Dshinks (talk) 20:06, 29 October 2020 (UTC)
@Dshinks: Honestly, the core MediaWiki search function is not great, and there's probably not a lot you can do about it... That said for your examples, WP:DAB for disambigs (particularly WP:ONEOTHER), WP:CAT for cats (ideally acyclic graph, hopefully not categories with few members), WP:AT and the linked naming conventions for titling. --Izno (talk) 02:55, 30 October 2020 (UTC)
To limit Google or Bing searches to the English Wikipedia, include this in your search string: site:en.wikipedia.org.
You can even do a Google search of Wikipedia from Wikipedia's own search box! For example, to use Google to search for pages with "geology" in the title, type this into WP's search box: google:geology+site:en.wikipedia.org
Not done: it's not clear what changes you want to be made. Please mention the specific changes in a "change X to Y" format and provide a reliable source if appropriate. Run n Fly (talk) 17:22, 10 May 2021 (UTC)
If I'm looking for a string which can contain Capital and Lower case greek letters, do I need to split them the way that I do in English or not. In other words, is the equivalent of the latin alphabet [A-Za-z]
I wish to search all articles (and exclude redirects) that start with Dr without the dot after it.
I found the option of searching by excluding redirects using Special:PrefixIndex. However there, I cannot enclose the search term with quotes. If it was allowed, I only had to enter "Dr " with a space. Without quotes, on giving Dr followed by a space, the search seems to be removing the trailing space, and returns all entries that start with "Dr". Note that I had asked the same question in July at WT:Special:PrefixIndex and not got a response. Jay(talk)05:34, 18 October 2021 (UTC)
@Jay: I added "Z" manually after seeing that "Dr" didn't work. A space is usually sorted before any other character but PrefixIndex apparently sorts space somewhere between "Z" and "a". PrimeHunter (talk) 20:12, 23 October 2021 (UTC)
searching for strings
"String" can have different meanings. For example, if I wish to search for placenames that include the string "borough" it's not clear from this information page how to do this, especially if one wants to make the results easier to navigate by excluding all occurrences of the word "borough". It'd be great if such a search could be executed by typing something like *borough -borough in the search box but it doesn't work. Dadge (talk) 04:07, 10 May 2020 (UTC)
My 'searching for strings' box has stopped working. For example I typed in France on an article about Reunion Island in the Indian Ocean and it doesn't come up with a single instance of the term although it's used multiple times. Any ideas how to get this very handy tool back? 7dmd7 (talk) 01:40, 9 December 2021 (UTC)
@7dmd7: It sounds like a feature of your browser or device, not Wikipedia. Is it there and does it work at other websites? Some search features have an option to specify case sensitive search. PrimeHunter (talk) 02:15, 9 December 2021 (UTC)
Statistics for the search terms used?
Is there any meta- or external source where we can get some statistics on the search terms used within Wikipedia? I know this is possible to do with Google, e.g. using [Google Trends, and I know it's possible to see the page views via the Stats links on the page histories. But is there something that shows search term usage within Wikipedia? Chumpih. (talk) 10:26, 28 December 2021 (UTC)
Range of Hebrew letters?
I can regex a range of Greek Letters without a problem. for example insource:/\('''[Α-Ω]+'''/ however, I am unable to figure out how to do the equivalent for Hebrew. None of the possibilities seem to work, replacing the A with an א and Z with ת nor doing it the other way around (so that the window thinks it is a single string reading RTL, nor even flipping the [ and ] seem to work, could someone please help me figure out the way to do a search for a parentheses, three quotes and a string of Hebrew letters?Naraht (talk) 14:17, 25 August 2021 (UTC)
I've been doing some maintenance using this whopper of a query; although it has some false positives, it's quite good at finding instances where someone just typed "citation needed" instead of using {{citation needed}}. Anyway, I made a huge AWB run and edited a bunch of the pages manually to fix the formatting. Yet if I run the search, I'll still see previews (with the misformatted original source text) for a bunch of articles I know I just fixed. I'm not entirely sure on how the search cache is built -- so this may be a completely impossible task -- but is there any way to tell the search engine that I demand a search of current article text, and not the cached search results from a couple hours ago? jp×g05:58, 21 November 2021 (UTC)
A proposed edit to the Medicago rigidula Wikipedia page under the Scientific classification. Division/phylum: Tracheophyta, Class: Magnoliopsida, Order: Fabales, Family: Leguminosae/Fabaceae.
Crystali18 (talk) 20:32, 12 April 2022 (UTC)
I believe search suggestions could have something to do with it. When I type "Mac" in the search box, Mac (computer) is the 10th suggestion. As I type more letters ("Macin", "Macint"), it disappears, replaced by articles for specific Macintosh models that receive far lower page views than the main article. The main article only appears again when I finish typing the full term ("Macintosh"), due to the redirect at Macintosh.
Could anything be done about this, short of moving the article back? Could the Macintosh redirect be treated as a normal page in the backend, so it shows up in search suggestions without needing an exact match? DFlhb (talk) 02:39, 21 January 2023 (UTC)
Unforgiving in namespace names
It occurs to me that the search is more strict in namespace name (prefix) than in title. Given a random typo (my favourite one): Tempalte, compare as namespace and as article title:
The autocomplete search results are different to search results, in that autocomplete for searches word order matter, but not for search results. Is there an option for autocomplete so that it gives me the same results as search results, including results with different word order? HudecEmil (talk) 07:27, 24 May 2023 (UTC)
Searching redirects
Is there a way to search redirects? Insource and Intitle seem to exclude redirects. I don't see anything in the docs about this. Dicklyon (talk) 00:23, 19 May 2023 (UTC)
Is there a reason why wildcard characters can't be used at the beginning of a search string? In other words, is there a way to search for articles whose titles end in a particular set of characters? Let's say, for example, I want to find every town whose name ends in "-borough". Zacwill (talk) 21:58, 17 October 2023 (UTC)
I don't have an answer for the first question, but I have for the second one: Unfortunately, MediaWiki's search engine doesn't support the regex metacharacters ^ and $ that are normally used to mark the start and end position of a string. But if you're familiar with PetScan and have a category, search string, pages that use a certain template or anything like that in mind, there's "Regexp filter" in the "Output" tab that you can use to find articles whose titles end in a particular set of characters by using ^ and $ in PetScan's "Regexp filter": ^.*Example$ (replace "Example" with your regex search string in that search string). For example, articles in "Category:Boroughs" whose name ends in "borough" or articles found with the search string "intitle:/borough/" whose name ends in "borough". --JAAqqO (talk) 22:48, 17 October 2023 (UTC)
That's very helpful, thank you. One more question: can I use this method to search Wikipedias besides the English Wikipedia? Zacwill (talk) 23:15, 17 October 2023 (UTC)
Thank you. That helps a lot. But intitle:/\"/ -incategory:"Articles with quotation marks in the title" finds both article titles and redirects. I don't want the redirects. — BarrelProof (talk) 03:44, 26 December 2023 (UTC)
In fact I think maybe that only finds redirects. I tried intitle:/\«/ and it didn't find the article at «O»«O», but it found a few redirects. Or maybe the problem has something to do with the timeout error I received. It says to "Try simplifying your regular expression to get complete results", but it is hard to conceive of a simpler regular expression than that. — BarrelProof (talk) 04:49, 26 December 2023 (UTC)
There's always PetScan that can filter out redirects from searches: use the Redirects field in the Page properties tab and the Search field in the Other sources tab. For example, here's a PetScan search for articles that 1) use the template {{Infobox film}}, 2) aren't redirects, 3) aren't in the Category:Articles with quotation marks in the title and 4) match the title regexp filter ^.*("|“|”|«|»).*$. "Try simplifying your regular expression to get complete results" – yeah, if you don't have a category, search string, pages that use a certain template or anything restrictive like that in mind (like the template "Infobox film" in my PetScan search), then—because of the limits of PetScan and MediaWiki's search engine—your best bet is to download the database and then use a tool like WP:AWB for searching. For example, based on the datadump enwiki-latest-pages-articles-multistream.xml.bz2 02-Dec-2023 03:27, I made a list that lists about 2,400 articles that, as of 2 December 2023, 1) have “, ”, ", « or » in their titles, 2) aren't in the category Category:Articles with quotation marks in the title and 3) aren't redirects. And here's a PetScan search that updates that list automatically. --JAAqqO (talk) 17:37, 27 December 2023 (UTC)
Unable to ID articles. Unable to search for text in the actual final pages.
this search for insource:/\<span style=\"border:thin solid black;\"\>WARNING\<\/span\>/ is failing.
I'm trying to find which of the pages linked to from https://en.wikipedia.org/wiki/User:RudolfoMD/sandbox2 do not display the boxed warning. I want all of them to display it. They all should, but some of the wikidata entries haven't been updated (I don't understand why; I'm an OpenRefine beginner.) It seems like search searches the wikitext, not the resulting HTML pages. Any ideas on how to do what I want - at the high medium or low level? (It seems that this is, de facto, a place where folks are asking questions about search problems and getting answers, so I'm going ahead.) RudolfoMD (talk) 05:52, 27 December 2023 (UTC)
So, are you looking for articles that don't show the input "Legal status: US: WARNING" in {{Infobox drug}} from the property legal status (medicine) (P3493)? If so, here are PetScan searches for the articles that are:
Thanks! I lost my initial comment thanking you. I was able to improve the situation somewhat using your searches. Both were useful! (how odd that an IP deleted this section!) RudolfoMD (talk) 06:07, 11 January 2024 (UTC)
Searching for strings only inside/outside refs?
Is there any way to limit a search to only inside or only outside of refs in an article? Inside seems like it *might* be possible using regex, outside, I don't really see how... Naraht (talk) 14:28, 12 March 2024 (UTC)
Just change the example word and regex in green. The insource:"Hello" or just simply Hello in the beginning of these search strings is used to restrict the search, so the search wouldn't time out and then give only partial results.
There's also a problem with that outside search: if a page has what you're searching for but also has the same matching search string inside a ref tag, the page won't be included in the search results.
I'd like to run a search for all articles to which I've added the {{IPAc-en}} template. Sometimes I left edit summaries saying so, sometimes not. How do I do this? Mac Dreamstate (talk) 21:19, 26 June 2024 (UTC)
I seem to get the same (and the expected) results for searches on All: "your base", all: "your base", All:"your base", all:"your base", namely between 100 and 250 results, with mainspace ones at the top. The space after : is not needed, and quotation marks would not be needed around a single-word term like EFFPR. I'm not sure what's causing your difficulties. What are the circumstances (browser versus mobile app, etc.)? — SMcCandlish☏¢ 😼 02:26, 8 July 2024 (UTC)
I'm using Google Chrome on PC (Windows 10), settings should all be default (as I use Wikipedia in incognito so any cookies reset every day and every time I close the browser).
I mean, I don't know about searching all: "your base" (that also gives me many results, though without the word all being highlighted (bolded) in the results) - but the thing with All: "EFFPR", is that it shows no results, presumably because for some reason it isn't searching for "EFFPR" in all namespaces like it does when the all: is in lowercase.
I get 247 results for all:EFFPR, all: EFFPR, all:"EFFPR", and all: "EFFPR"; all identical (and expected) behavior. However, All: EFFPR produces bogus results of 259,538, matching on strings like effort. Meanwhile, All:EFFPR, All:"EFFPR", and All: "EFFPR" each produce no hits. So, there is clearly an interpretation consistency problem of some kind here. The documentation should be changed to say to use all: not All:, and this may affect other such keywords as well. I would think a phab ticket also needs to be opened about this, since case-sensitivity of these search keywords is not expected behavior. Nor is All: EFFPR producing bogus matches for effort and other strings that just contain eff... substrings. I have no idea where it got the idea to do a lower-casing substring match. PS: It's weird that All: "your base" produces 139 hits, All:"your base" same (seemingly all confined to mainspace, and with a few substring matches such as just all (presumably near your or base; but all:"your base" and all: "your base" produce 2,529 hits (sorted by namespace, though also wandering eventually into substring matches like "when your base assumption is"). I don't really know what's going on here. But clearly a capitalized All: either fails to find all the applicable results (excluding all but mainspace, and not findinal all occurrences in mainspace), or fails to find any at all, or falsely matches random word-fragment substrings; while the worst that can be said about the all: version is that it will eventually start ignoring the "..." constraint and will include substring matches (though seemingly only whole-word ones, not word-fragment ones) instead of only the exact phrase. — SMcCandlish☏¢ 😼 03:11, 8 July 2024 (UTC)
What you described matches what happens for me, yeah. An observation:
- Special:Search/All: "your base" -> finds 139 results (all in article space), second result is Tatsuya Uemura (category All stub articles);
The first 2 searches being identical makes me think that the word All is being included as a search term, that's why it fails to find some results, because they don't have the word all: 145 - 6 = 139.
About this "hyper-anonymity", I have to ask "To what end?" Accounts here are pseudonymous, and you need not use one you create for more than whatever purpose an account is needed, not even more than once. You could just create an account named "User:2804-etc." and use it for no purpose other than this sort of thing. I might get around to opening a phab ticket, but I have about 100+ projects of all sorts going on at once (only a small fraction of them to do with WP). — SMcCandlish☏¢ 😼 17:33, 19 July 2024 (UTC)