Search is case insensitive, that is, upper and lowercase is ignored.
Search "folds" character families, matching similar-looking letters across alphabets, to match foreign terms. So, you don't have to type in diacritical
letters, and your terms will still match. For example, Citroen
will match Citroën
, and Aeroskobing
Characters that are not numbers or letters (punctuation marks, brackets and slashes, math and other symbols) are generally ignored. For example, Credit (finance)
will return articles with the words credit
, ignoring the parentheses, unless an article with exact title Credit (finance)
exists. Similarly, a search for the string |LT|
between two pipe symbols) will only return articles with lt
. In order to search for terms that contain non-alphanumeric characters, a regex search
must be used instead (using the \ escape character
if required), for example insource:/\|LT\|/
will successfully return all instances of |LT|
The source text is what is searched, which is not always what is displayed on the screen. This distinction is relevant for piped links
, for interlanguage links
(to find links to Chinese articles, search for zh
, not for Zhongwen
), special characters
is coded as ê
it is found searching for ecirc
The default search domain is the article space, but any namespace may be specified in a query.
At the search results page, any number of namespaces
can be specified, and users
can keep those namespaces as their own default search domain. Partial namespace searches can be made by specifying the initial letters of a page name.
The use of spaces is, in general, intuitive. Unnecessary spaces, and all non-alphanumeric characters except " are ignored, which makes for flexibility; it is simplest and best to avoid typing unnecessary spaces, although the tolerance for grey space simplifies copying and pasting search terms without the need for cleanup. For example, credit card is obviously reasonable; copying and pasting [[Credit(?!)card]] is equivalent and convenient; "credit card""payment card" is actually correct and minimal, but "credit card" "payment card" is a sensible equivalent.
Any of the following characters will be treated as a space character
. The double quotation mark "
is not one of these characters, because it has the special purpose of specifying an exact phrase search, and -
can be used to exclude results if either occurs at the beginning of a word or phrase (see more below). We use the term grey-space
instead of whitespace
here to include the space character itself and all these characters. Multiple [grey-]spaces are equivalent to a single space.
Grey-space is ignored between the words of exact-phrase searches, between adjacent items in the query, and in starting characters of the search box query. All filters can have grey-space between them without affecting search results. Most operators, such as intitle: and incategory:, ignore unnecessary spaces, or grey-space, after the colon.
Where spaces are significant: single search terms cannot have embedded spaces; work space, "work space", and workspace are all different. The particular keywords prefix and insource must be followed immediately by a colon:and their arguments, without intervening [grey-]spaces.
Double quotes for exact phrase search
can be matched by enclosing it in double quotes, "like this"
. Double quotes can define a single search term that contains spaces. For example, "holly dolly"
where the space is quoted as a character, differs much from holly dolly
where the space is interpreted as a logical AND.
Suffixed tilde character for fuzzy search
is requested by suffixing a tilde (~
) like this~
, with results like "thus" and "thins". It covers any two character-changes for any character except the first: it returns addition, exchange, or subtraction. This search technique is sometimes called a "sounds-like" search. For example, searching for charlie~ parker~
returns Charlie Parker
, Charles Palmer
, Charley Parks
Prefixed tilde character for forced search
To force a search rather than navigate directly to a matching page, include a tilde character ~
anywhere in the query. It always takes you to the search results page, never jumping to a single title. For example, the misspelling similiar
to the Similarity
article, but prefixing a tilde, ~similiar
, lists pages containing that misspelling.
Prefixed hyphen or exclamation point for exclusion
Pages matching a search term can be excluded by prefixing an exclamation point (!) or a hyphen or dash (-
) to the term. This is the logical NOT
. For example, credit card -"credit card"
finds all articles with "credit" and "card" except
those with the phrase "credit card".
The two wildcard characters are * and \?, and both can come in the middle or end of a word. The escaped question mark stands for one character and the star stands for any number of characters. Because many users ask questions when searching, question marks are ignored by default, and the escaped question mark (\?) must be used for a wildcard.
For non-alphabetic characters, regex expressions are needed. See #insource:
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.
Search terms are implicitly joined by AND
. For example "credit card" housecat
searches for pages containing both "credit card" and "housecat". An OR
operator is supported, but will only give intuitive results (corresponding to logical disjunction
) if all
search terms are separated by OR
(e.g. red OR green OR blue
has the expected behaviour, but red OR green blue
does not). OR
also does not behave predictably with special keywords (like intitle:
) or namespaces. See mw:Help:CirrusSearch/Logical operators
for a more detailed explanation.
Parameters function as name filters, each followed by the search term it operates on. Their search term may be a word or a phrase. The main parameters are namespace:, intitle:, insource:, incategory:, and prefix:. ("namespace" as used here isn't literal – use the name of the actual namespace desired).
"prefix:" differs from the other parameters in that it can only be used at the end of a search string.
A single "namespace:" filter can go first, and a single "prefix" filter can go last.
A reader searching for articles from the search box need know nothing about namespaces, so the default user preferences
are set to search only in article space; but an advancing editor can reset the default search-space preference for the current search to a particular namespace, or "all" by prefixing the search with "all:" or the namespace name followed by a colon.
Given only at the beginning
of the query, a namespace name followed by a colon limits search results to that namespace. It is a filter without a query string. Namespace aliases
, like "WP" for "Wikipedia", are accepted.
Prefixing "All:" to a search string, searches all namespaces, and prioritizes mainspace matches to the top.
Using the lower-case "all:" version also searches all namespaces but does not prioritize the results by namespace.
and redirects can be searched with "intitle:query
". The search results highlight occurrences in both the title and page content. Multiple "intitle" filters may be used to search for words in titles regardless of order, or possible in different titles (i.e., redirects) for the same article. Regular expressions
can be used with "intitle:/regexp
/" or the case insensitive "intitle:/regexp
/i". See more in the insource
This can find template arguments, URLs, links, html, etc. It has two forms, one is an indexed search, and the other is regex
Like word searches and exact-phrase searches, non-alphanumeric characters are ignored, and proximity and fuzziness are options.
These are regular expressions. They use a lot of processing power, so we can only allow a few at a time on the search cluster, but they are very powerful. The version with the extra i runs the expression case-insensitive, and is even less efficient. Regex searches are likely to time out unless you further limit the search in some way, such as by including another parameter or a search term outside of the insource component of the search string. (For example, X* intitle:/X/ to restrict the search to initial position.) For more details, see mw:Help:CirrusSearch#Regular expression searches.
Given as "incategory:category
", where category
is the page name
of a category page, it lists pages with
in their wikitext. (Editors searching in namespaces other than mainspace will need to know the limitations
these search results may contain.) Space characters in a page name can be replaced with an underscore instead of using double quotes; either way works, and even both at once works (but not on Commons). "Incategory:" will also return pages in the adjacent subcategory; see for example, "category: incategory:History
". Multiple "incategory" filters may be applied. A more graphical alternative to a single filter is at Special:CategoryTree
. Because categories are important structures for searching
for related articles, any use of this prefix is particularly effective for searching. For more on using the categories themselves to find articles, see Wikipedia:FAQ/Categories
Also deepcat:, this allows you to search in a category and all its subcategories. The depth of the tree is currently limited to 5 levels, and the overall number of categories is limited to 256.
deepcat:"Musicals by topic"
" patterns only the beginning characters
of a page name
. Because the "beginning" characters can, if you need, go on to include the characters all the way to the end of the page name, prefix must include spaces, since page names often include spaces. For this reason prefix:
must only ever be given at the last part
of a search box query, and next character after the colon cannot be a space. Prefix does not search for partial namespace names, but requires at least a full namespace name to start to find pages, but prefix:
also recognizes an alias
of a namespace, and it recognizes redirects (or shortcut). Prefix is the most widely used and powerful filter as it can mimic the namespace filter, and because intitle:
cannot easily target a single page, even together with other filters. Special:PrefixIndex
is a MediaWiki, graphical, version, using only prefix:
to find pages.
"linksto:page name" searches in pages that link to the given page. Can be used negatively by prefixing a hyphen, which will return pages that do not link to the given page. Unlike with some other keywords, the page name is case-sensitive.
This only works right in the search box of the search page or search results page.
This limits searches to subpages of the specified page. You can also negate the subpages from a search by preceding subpageof: with a hyphen. Note that articles on Wikipedia generally don't have subpages, but the pages of the other namespaces do. Therefore, use the namespace parameter also. Here are some examples:
To look at all the subpages of a user, try this:
To make sure Articles for deletion pages do not show up in the results of a Wikipedia namespace search, try this:
Wikipedia:"Hi there" -subpageof:"Articles for deletion"
That'll show pages from the Wikipedia namespace with "Hi there" in them, and the list of results will not be cluttered with any Articles for deletion debates (many thousands of which are in the Wikipedia namespace). Notice the use of the hyphen (that makes it mean "not subpages of".)
This finds pages that use the specified template. Input the canonical page name to find all usage of the template, but use any of its redirect page names finds just that naming. Namespace aliases are accepted, capitalization is entirely ignored, and redirects are found, all in one name-search.
This is more thorough than insource:, in that it will find meta-templates (templates called by another template). Meta-templates don't show up in the local page's wikitext.
Examples of hastemplate: usage:
hastemplate:"Article for deletion/dated"
This lets you find all the articles being considered for deletion.
intitle:"Outline of " -hastemplate:"Outline footer"
This lets you find Wikipedia outlines that are lacking the outline footer template. (Notice the use of the hyphen to indicate "NOT").
This keyword allows filtering search results by topic. For possible topics see mw:Help:CirrusSearch/articletopic
. E.g. articletopic:books
will filter the search results to articles about books. articletopic:books|films
will filter to articles about books or films. articletopic:books articletopic:films
will filter to articles which are about both books and films.
Only mainspace articles belong to topics. Unlike other filters, articletopic also does page weighting: articles which are a stronger match for a topic will be higher in the search results (while articles which aren't about that subject at all will be removed from the result set completely).
Topic models are derived via machine learning from ORES
. Any given article receives a score on dozens of different topics, and therefore may appear under different keywords. For instance, the article on Albert Einstein may appear as a "physics" article and a "biography" article. Topic-related search data is updated weekly, so recently created articles might not show up in topic-based search queries.