幫助:扩展:解析器函數

From MediaWiki.org
Jump to navigation Jump to search
This page is a translated version of the page Help:Extension:ParserFunctions and the translation is 74% complete.
Other languages:
Bahasa Indonesia • ‎Deutsch • ‎English • ‎Lëtzebuergesch • ‎Nederlands • ‎Tiếng Việt • ‎Türkçe • ‎Zazaki • ‎español • ‎français • ‎galego • ‎italiano • ‎magyar • ‎polski • ‎português • ‎português do Brasil • ‎slovenščina • ‎suomi • ‎svenska • ‎български • ‎русский • ‎українська • ‎հայերեն • ‎العربية • ‎فارسی • ‎پښتو • ‎मराठी • ‎বাংলা • ‎தமிழ் • ‎ไทย • ‎中文 • ‎日本語 • ‎한국어

扩展:解析器函数 扩展提供11个附加解析器函数以补充MediaWiki固有的“魔术字”。 (它有可能会配置到增加的解析器函数以支持运行:这些字符串函数已在其他文档页面记载。) 所有的解析器函数由下列的表格提供。

{{#函数名: 参数1 | 参数2 | 参数3... }}

#expr

类型 运算符号
分组(括弧) ( )
数字 1234.5   e (2.718)   pi (3.142)
二进制操作者 e   一元 +,-
因为用于$ifeq和$switch的PHP用整数类型比较两个数,它会正确返回预期结果。然而用$expr比较相同的两个数时: not ceil trunc floor abs exp ln sin cos tan acos asin atan
二进制 ^
* / div mod
+ -
取整 round
逻辑 = != <> > < >= <=
and
or

这个函数计算并返回数学表达式的结果。 这个函数也通过mw.ext.ParserFunctions.expr函数中在扩展:Scribunto 有效。

{{#expr: 表达式 }}

右表依优先级列出了支持的运算符,运算符的详细说明请见Help:Calculation。运算结果的精度和格式受服务器操作系统及语言设置影响可能存在差异。

使用布尔代数时,0表示false,其他任何非0数值(无论正负)均表示true

{{#expr: 1 and -1 }}1
{{#expr: 1 and 0 }}0
{{#expr: 1 or -1 }}1
{{#expr: -1 or 0 }}1
{{#expr: 0 or 0 }}0

空表达式返回空值,错误的表达式返回错误信息,使用#iferror函数检查错误:

{{#expr: }}
{{#expr: 1+ }}Expression error: Missing operand for +.
{{#expr: 1 = }}Expression error: Missing operand for =.
{{#expr: 1 foo 2 }}Expression error: Unrecognized word "foo".

置于数字前的加减号会被视为正负号而不会报错:

{{#expr: +1 }}1
{{#expr: -1 }}-1
{{#expr: + 1 }}1
{{#expr: - 1 }}-1

注意:使用魔术字输出的数值时,首先要将其格式化以去除逗号,获得纯数字。例如:{{NUMBEROFUSERS}}输出17,227,131,而期望的输出值是17227131,使用{{formatnum :{{NUMBEROFUSERS}}|R}}便可以实现。对于某些语言,格式化数字尤为重要。例如孟加拉语中,{{NUMBEROFUSERS}}的输出结果是৩০,০৬১。

{{#expr:{{NUMBEROFUSERS}}+100}} Expression error: Unrecognized punctuation character ",".
{{#expr:{{formatnum:{{NUMBEROFUSERS}}|R}}+100}}17227231
警告 警告: 运算符mod會對第二參數的某些值給出錯誤的结果:
{{#expr: 123 mod (2^64-1)}}Division by zero. (輸出一個空值;應該是123)
如果你想要进行基于日期时间的计算(如检测当前日期时间是否超过指定日期时间),首先用{{#time: xNU }}将日期时间转换为距1970年1月1日的秒数,再对秒数进行加减运算。

四舍五入

a round b是指将a四舍五入至(1/10)b的整数倍。

向上、向下取整分别使用ceilfloor

案例 结果 凑整方法
{{#expr: 1/3 round 5 }} 0.33333 最终的数字小于5,所以没有明显的取整发生
{{#expr: 1/6 round 5 }} 0.16667 最终的数字大于5,所以被向上取整
{{#expr: 8.99999/9 round 5 }} 1 同理,最后一位数因后一位而加1,进而引起了前面的进位
{{#expr: 1234.5678 round -2 }} 1200 舍入至最近的整百数,负值在小数点左边。
{{#expr: 1234.5678 round 2 }} 1234.57 舍入至最近的整1/100数,因为正值会舍入至小数点右边。
{{#expr: 1234.5678 round 2.3 }} 1234.57 舍入指数的小数点在舍入结果中不起作用
{{#expr: trunc 1234.5678 }} 1234 小数部分被全部舍弃(截断)
四舍五入至最近的整数
{{#expr: 1/3 round 0 }} 0 向下取整至0
{{#expr: 1/2 round 0 }} 1 向上取整至1
{{#expr: 3/4 round 0 }} 1 向上取整至1
{{#expr: -1/3 round 0 }} -0 向上取整至0
{{#expr: -1/2 round 0 }} -1 向下取整至-1
{{#expr: -3/4 round 0 }} -1 向下取整至-1
使用ceil和floor向上或向下取整
{{#expr: ceil(1/3) }} 1 向上取整至1
{{#expr: floor(1/3) }} 0 向下取整至0
{{#expr: ceil(-1/3) }} -0 向上取整至0
{{#expr: floor(-1/3) }} -1 向下取整至-1
{{#expr: ceil 1/3 }} 0.33333333333333 没有取整,因为1本来就是整数
警告 警告: 注意会被解析为(ceil 1)/3,而不是你想象中的ceil(1/3)

字符串

表达式函数只用于仅含有数字和运算符号的表达式,而字母字符串和其它符号不在此列。若需要比较字符串,可以用#ifeq替代。

{{#expr: "a" = "a" }}Expression error: Unrecognized punctuation character """.
{{#expr: a = a }}Expression error: Unrecognized word "a".
{{#ifeq: a | a | 1 | 0 }}1

#if

这个函数判断一个字符串是否为空。只包含空格的字符串被视为空字符串。

{{#if: 测试字符串 | 当测试字符串不是空值时的值 | 当测试字符串是空值(或只包括空格)时的值 }}
{{#if: 参数1 | 参数2 | 参数3 }}

这个函数首先判断参数1是否为空。如果参数1不为空,则输出参数2。如果参数1为空或只包含空白内容(空格、换行等),则输出参数3。

{{#if: | yes | no}}no
{{#if: string | yes | no}}yes
{{#if:      | yes | no}}no
{{#if:


| yes | no}}
no

字符串内容会被处理为纯文本,因此数学表达式不会被计算:

{{#if: 1==2 | yes | no }}yes
{{#if: 0 | yes | no }}yes

最后一个参数(false)可以被忽略:

{{#if: foo | yes }} yes
{{#if: | yes }}
{{#if: foo | | no}}

该函数可嵌套。为了做到这一点,嵌套内层的#if函数需要按规则完整填写所有参数以符合#if的参数形式。一般至多可嵌套7层,该数也可能为wiki自身限制和内存限制而有变动。

{{#if: 测试字符串 | 字符串非空时的值 | {{#if: 测试字符串 | 字符串非空时的值 | 字符串为空(或只有空格)的值 }} }}

你可以在#if条件句中以字符串变量代替测试字符串。在这里,你需要在变量名的后面加上| (分隔符)以正确引用字符串。

{{#if:{{{1|}}}|你在变量1里面输入了文本|变量1里面没有文本 }}

在这里 Help:Parser functions in templates 获得更多关于解释器函数中变量的相关例子。

#ifeq

这个函数判断两个輸入字符串是否相同,並根據結果輸出兩個字符串的其中一個。如果需要更多的比較和輸出字符串,請考慮使用#switch

{{#ifeq: string 1 | string 2 | value if identical | value if different }}

如果两个字符串均为数字,则函数会进行数值的比较:

{{#ifeq: 01 | 1 | equal | not equal}}equal
{{#ifeq: 0 | -0 | equal | not equal}}equal
{{#ifeq: 1e3 | 1000 | equal | not equal}}equal
{{#ifeq: {{#expr:10^3}} | 1000 | equal | not equal}}equal

否则函数会进行文本的比较(大小写敏感):

{{#ifeq: foo | bar | equal | not equal}}not equal
{{#ifeq: foo | Foo | equal | not equal}}not equal
{{#ifeq: "01" | "1" | equal | not equal}}not equal  (对比上面没有引号的例子)
{{#ifeq: 10^3 | 1000 | equal | not equal}}not equal  (对比上面带有#expr的例子,會先回傳一個有效的整數)

作为例子,考虑一个已经存在着的模板 Template:Timer,该模板利用解释器来返回事先设定对应的时间(shortlong)。该模板将第一个输入的参数(如果有)作为变量,与字符串"short"(该字符串不存在既定规则的与其他类型的量的转换)纯粹的做比较。模板的定义如下:

{{#ifeq: {{{1|}}} | short | 20 | 40 }}

会产生如下结果:

{{timer|short}}20
{{timer|20}}40
{{timer}}40
警告 警告: 除了下面这种情况:所输入表达式为空或者是一个错误表达式(空串会返回一条错误信息,而它不等于0,所以在后者我们会得到"value if true")。
{{#ifeq: 12345678901234567 | 12345678901234568 | equal | not equal}}not equal
{{#switch: 12345678901234567 | 12345678901234568 = equal | not equal}}not equal

因为用于#ifeq#switch的PHP用整数类型比较两个数,它会正确返回预期结果。然而用#ifexpr比较相同的两个数时:

{{#ifexpr: 12345678901234567 = 12345678901234568 | equal | not equal}}equal
由于两个数存在不相同的数位,相等比较的结果显然是“错误的”。造成#ifexpr出错行为的原因是MediaWiki将两个数依字面表达理解成了浮点数,而对于像这样的大数,转换为浮点数由于精度误差会存储为一样的数值,导致出错。

警告 警告: 如果比较参数使用了解释器函数,任何的解释器标记和其它的解释器函数都一定会被一个临时生成的独一无二的标识码所替代。这会影响如下列的比较:
{{#ifeq: <nowiki>foo</nowiki> | <nowiki>foo</nowiki> | equal | not equal}}not equal
{{#ifeq: <math>foo</math> | <math>foo</math> | equal | not equal}}not equal
{{#ifeq: {{#tag:math|foo}} | {{#tag:math|foo}} | equal | not equal}}not equal
{{#ifeq: [[foo]] | [[foo]] | equal | not equal}}equal
如果被比较字符串从所使用的模板到模板内使用的参数和标记都是相同的,那么这种情况下会返回真值,但如果所使用模板名不同,即便内容是完全一致的,判断也会返回假值。
警告 警告: 随网站配置的变化,含有页面名字的魔术字的字符串的字面比较可能会发生错误。 比如说,{{FULLPAGENAME}}这一魔术字随wiki的变化,可能在返回时使开头字母变成大写,也会用下划线替代空格。

To work around this, apply the magic word to both parameters:

{{#ifeq: {{FULLPAGENAME: L'Aquila}} | {{FULLPAGENAME}} | equal | not equal}}equal

#iferror

这个函数以一个字符串为输入,然后在两条预选结果中取其一而返回。函数认定为true当且仅当作为输入参数的字符串在解释时返回的结果中含有一个class="error"的HTML对象,而其中可用来认定结果的包括其他的解释器函数(比如#expr#time#rel2abs)的错误,模板错误(比如模板循环和模板递归),还有一些其它的解释器“软错误”。

{{#iferror: test string | value if error | value if correct }}

待返回字符串参数可以省略。若省略"correct"(正确)字符串参数,则在"test string"(测试字符串)不出错的情况下函数将返回被测字符串本身。若省略"error"(错误)字符串参数,则函数将在被测字符串出错时返回空字符串:

{{#iferror: {{#expr: 1 + 2 }} | error | correct }}correct
{{#iferror: {{#expr: 1 + X }} | error | correct }}error
{{#iferror: {{#expr: 1 + 2 }} | error }}3
{{#iferror: {{#expr: 1 + X }} | error }}error
{{#iferror: {{#expr: 1 + 2 }} }}3
{{#iferror: {{#expr: 1 + X }} }}
{{#iferror: {{#expr: . }} | error | correct }}correct
{{#iferror: <strong class="error">a</strong> | error | correct }}error

#ifexpr

此函数会判断数学表达式并根据其布尔值结果返回对应的字符串:

{{#ifexpr: expression | value if true | value if false }}

这里"expression"输入串将原封不动的作为上面#expr的参数进行调用,且表达式运算符是通用的,返回值也将作为布尔表达式进行处理。

输入表达式为空时将视为false

{{#ifexpr: | yes | no}}no

如上面所提,0将视为false,非零值将视为true,因此这个函数与下面的仅用#ifeq#expr的表述等价:

{{#ifeq: {{#expr: expression }} | 0 | value if false | value if true }}

除了下面这种情况:所输入表达式为空或者是一个错误表达式(空串会返回一条错误信息,而它不等于0,所以在后者我们会得到"value if true")。

{{#ifexpr: = | yes | no }} Expression error: Unexpected = operator.

相对的

{{#ifeq: {{#expr: = }} | 0 | no | yes }} yes

两个待选返回值都不是必填的。当表达式判断对应的分支返回值被省略时,函数不会给出输出:

{{#ifexpr: 1 > 0 | yes }}yes
{{#ifexpr: 1 < 0 | yes }}
{{#ifexpr: 0 = 0 | yes }} yes
{{#ifexpr: 1 > 0 | | no}}
{{#ifexpr: 1 < 0 | | no}} no
{{#ifexpr: 1 > 0 }}

#ifexist

这个函数将一组字符串作为输入,并翻译成页面标题,然后根据在本地wiki上是否存在该页面而返回对应的值。

{{#ifexist: page title | value if exists | value if doesn't exist }}

只要页面存在就会判定为true(真值),即便那个页面看上去是空白的(比如像是分类链接或者是魔术字解释页之类的却不包含任何可视内容的页面),或者是重定向页,或者它就是空白页。当且仅当页面是红链时判定为false(假值),包括那些曾经存在却被删除的页面。

{{#ifexist: Help:Extension:ParserFunctions/zh | exists | doesn't exist }}exists
{{#ifexist: XXHelp:Extension:ParserFunctions/zhXX | exists | doesn't exist }}doesn't exist

函数在系统消息对应模块被安装时认定为true,对特殊页面的判定则取决于本地软件自身。

{{#ifexist: Special:Watchlist | exists | doesn't exist }}exists
{{#ifexist: Special:CheckUser | exists | doesn't exist }}exists (因为Checkuser 扩展已经安装于此wiki)
{{#ifexist: MediaWiki:Copyright | exists | doesn't exist }}exists 最后一个参数(false)可以被忽略:

If a page checks a target using #ifexist:, then that page will appear in the Special:WhatLinksHere list for the target page. So if the code {{#ifexist:Foo }} were included live on this page (Help:Extension:ParserFunctions/zh), Special:WhatLinksHere/Foo will list Help:Extension:ParserFunctions/zh.

On wikis using a shared media repository, #ifexist: can be used to check if a file has been uploaded to the repository but not to the wiki itself:

{{#ifexist: File:Example.png | exists | doesn't exist }}doesn't exist
{{#ifexist: Image:Example.png | exists | doesn't exist }}doesn't exist
{{#ifexist: Media:Example.png | exists | doesn't exist }}exists

If a local description page has been created for the file, the result is exists for all of the above.

#ifexist: 不会对维基内链起作用。

ifexist 限制

#ifexist: is considered an "expensive parser function"; only a limited number of which can be included on any one page (including functions inside transcluded templates). When this limit is exceeded, any further #ifexist: functions automatically return false, whether the target page exists or not, and the page is categorized into Category:Pages with too many expensive parser function calls. The name of the tracking category may vary depending on the content language of your wiki.

For some use cases it is possible to emulate the ifexist effect with css, by using the selectors a.new (to select links to unexisting pages) or a:not(.new) (to select links to existing pages). Furthermore, since the number of expensive parser functions that can be used on a single page is controlled by $wgExpensiveParserFunctionLimit , one can also increase the limit in LocalSettings.php if needed.

ifexist and wanted pages

A page that does not exist and is tested for using #ifexist will end up on the Wanted Pages.

#rel2abs

This function converts a relative file path into an absolute filepath.

{{#rel2abs: path }}
{{#rel2abs: path | base path }}

Within the path input, the following syntax is valid:

  • . → the current level
  • .. → "go up one level"
  • /foo → "go down one level into the subdirectory /foo"

If the base path is not specified, the full page name of the page will be used instead:

{{#rel2abs: /quok | Help:Foo/bar/baz }}Help:Foo/bar/baz/quok
{{#rel2abs: ./quok | Help:Foo/bar/baz }}Help:Foo/bar/baz/quok
{{#rel2abs: ../quok | Help:Foo/bar/baz }}Help:Foo/bar/quok
{{#rel2abs: ../. | Help:Foo/bar/baz }}Help:Foo/bar

Invalid syntax, such as /. or /./, is ignored. Since no more than two consecutive full stops are permitted, sequences such as these can be used to separate successive statements:

{{#rel2abs: ../quok/. | Help:Foo/bar/baz }}Help:Foo/bar/quok
{{#rel2abs: ../../quok | Help:Foo/bar/baz }}Help:Foo/quok
{{#rel2abs: ../../../quok | Help:Foo/bar/baz }}quok
{{#rel2abs: ../../../../quok | Help:Foo/bar/baz }}Error: Invalid depth in path: "Help:Foo/bar/baz/../../../../quok" (tried to access a node above the root node).

#switch

This function compares one input value against several test cases, returning an associated string if a match is found.

{{#switch: comparison string
 | case = result
 | case = result
 | ...
 | case = result
 | default result
}}

例如:

{{#switch: baz | foo = Foo | baz = Baz | Bar }} Baz
{{#switch: foo | foo = Foo | baz = Baz | Bar }} Foo
{{#switch: zzz | foo = Foo | baz = Baz | Bar }} Bar

#switch with partial transclusion tags can effect a configuration file that enables an editor unfamiliar with template coding to view and edit configurable elements.

默认

The default result is returned if no case string matches the comparison string:

{{#switch: test | foo = Foo | baz = Baz | Bar }} Bar

In this syntax, the default result must be the last parameter and must not contain a raw equals sign.

{{#switch: test | Bar | foo = Foo | baz = Baz }} →
{{#switch: test | foo = Foo | baz = Baz | B=ar }} →

Alternatively, the default result may be explicitly declared with a case string of "#default".

{{#switch: comparison string
 | case = result
 | case = result
 | ...
 | case = result
 | #default = default result
}}

Default results declared in this way may be placed anywhere within the function:

{{#switch: test | foo = Foo | #default = Bar | baz = Baz }} Bar

If the default parameter is omitted and no match is made, no result is returned:

{{#switch: test | foo = Foo | baz = Baz }}

组状态

It is possible to have 'fall through' values, where several case strings return the same result string. This minimizes duplication.

{{#switch: comparison string
 | case1 = result1
 | case2 
 | case3 
 | case4 = result234
 | case5 = result5
 | case6 
 | case7 = result67
 | #default = default result
}}

Here cases 2, 3 and 4 all return result234; cases 6 and 7 both return result67

比较行为

As with #ifeq, the comparison is made numerically if both the comparison string and the case string being tested are numeric; or as a case-sensitive string otherwise:

{{#switch: 0 + 1 | 1 = one | 2 = two | three}} → three
{{#switch: {{#expr: 0 + 1}} | 1 = one | 2 = two | three}} → one
{{#switch: a | a = A | b = B | C}} → A
{{#switch: A | a = A | b = B | C}} → C

A case string may be empty:

{{#switch: | = Nothing | foo = Foo | Something }}Nothing

Once a match is found, subsequent cases are ignored:

{{#switch: b | f = Foo | b = Bar | b = Baz | }}Bar
警告 警告: Numerical comparisons with #switch and #ifeq are not equivalent to comparisons in expressions (see also above):
{{#switch: 12345678901234567 | 12345678901234568 = A | B}} → B
{{#ifexpr: 12345678901234567 = 12345678901234568 | A | B}} → A

原始等号

"Case" strings cannot contain raw equals signs. To work around this, create a template {{=}} containing a single equals sign: =, or replace equals sign with html code &#61;.

如果比较参数使用了解释器函数,任何的解释器标记和其它的解释器函数都一定会被一个临时生成的独一无二的Template:$1所替代。这会影响如下列的比较:

{{#switch: 1=2
 | 1=2 = raw
 | 1<nowiki>=</nowiki>2 = nowiki
 | 1{{=}}2 = template
 | default
}}template


{{#switch: 1=2
 | 1&#61;2 = html
 | default
}}html

注意 注意: For a simple real life example of the use of this function, check Template:NBA color. Two complex examples can be found at 模板:扩展 and w:Template:BOTREQ.

替换#ifeq

#switch can be used to reduce expansion depth.

例如:

  • {{#switch:{{{1}}} |condition1=branch1 |condition2=branch2 |condition3=branch3 |branch4}}

等效于

  • {{#ifeq:{{{1}}}|condition1 |branch1 |{{#ifeq:{{{1}}}|condition2 |branch2 |{{#ifeq:{{{1}}}|condition3 |branch3 |branch4}}}}}}

i.e. deep nesting, linear:

{{#ifeq:{{{1}}}|condition1
  |<!--then-->branch1
  |<!--else-->{{#ifeq:{{{1}}}|condition2
                |<!--then-->branch2
                |<!--else-->{{#ifeq:{{{1}}}|condition3
                              |<!--then-->branch3
                              |<!--else-->branch4}}}}}}

On the other hand, the switch replacement could be complicated/impractical for IFs nested in both branches (shown with alternatives of indentation, indented on both sides), making full symmetrical tree:

{{#ifeq:{{{1}}}|condition1
 |<!--then-->branch1t{{
  #ifeq:{{{1}}}|condition2
   |<!--then-->branch1t2t{{#ifeq:{{{1}}}|condition4|<!--then-->branch1t2t4t|<!--else-->branch1t2t4e}}
   |<!--else-->branch1t2e{{#ifeq:{{{1}}}|condition5|<!--then-->branch1t2e5t|<!--else-->branch1t2e5e}}
  }}
 |<!--else-->branch1e{{#ifeq:{{{1}}}|condition3
   |<!--then-->branch1e3t{{#ifeq:{{{1}}}|condition6|branch1e3t6t|branch1e3t6e}}
   |<!--else-->branch1e3e{{
    #ifeq:{{{1}}}|condition7
     |branch1e3e7t
     |branch1e3e7t
    }}
  }}
}}

#time

代码 描述 当前输出
刷新此页面的缓存以更新)
Y 4位数年。 2019
y 2位数年。 19
L 如果是闰年则为1,否则为0。 0
o[note 1] 指定周的ISO-8601年。[note 2] 2019[note 3]
  1. 需要 PHP 5.1.0 及更新版本和 rev:45208.
  2. 这拥有与Y相同的值,除了当ISO周数(W)属于去年或明年时,会改用那一年。
  3. 如果未满足注释1,将输出文字“o”。
n 月指数,非零填充。 11
m 月指数,零填充。 11
M 月名的缩写,使用网站语言。 11月
F 使用网站语言的完整月名。 11月
xg Output the full month name in the genitive form for site languages that distinguish between genitive and nominative forms. This option is useful for many Slavic languages like Polish, Russian, Belarusian, Czech, Slovak, Slovene, Ukrainian, etc. 波兰语:
{{#time:F Y|June 2010|pl}} → czerwiec 2010
会产生如下结果:
{{#time:d xg Y|20 June 2010|pl}} → 20 czerwca 2010
(属格)
Day of the month or the year
j 月内一日,非零填充。 12
d 月内一日,零填充。 12
z Day of the year (January 1 = 0).
注意 注意: To get the ISO day of the year add 1.
315
Week and day of the week
W ISO 8601周数,零填充。 46
N ISO 8601 一周中的某天 (Monday = 1, Sunday = 7). 2
w Number of the day of the week (Sunday = 0, Saturday = 6). 2
D An abbreviation for the day of the week. Rarely internationalized.
l The full weekday name. Rarely internationalized. 星期二
小时
a "am" during the morning (00:00:00 → 11:59:59), "pm" otherwise (12:00:00 → 23:59:59). pm
A Uppercase version of a above. PM
g Hour in 12-hour format, not zero-padded. 4
h Hour in 12-hour format, zero-padded. 04
G Hour in 24-hour format, not zero-padded. 16
H Hour in 24-hour format, zero-padded. 16
分钟和秒钟
i Minutes past the hour, zero-padded. 08
s Seconds past the minute, zero-padded. 48
U Unix时间。 Seconds since January 1 1970 00:00:00 GMT. 1573574928
时区(截止1.22wmf2
e 时区标识符。 UTC
I Whether or not the date is in daylight savings time. 0
O 与格林尼治时间(GMT)的差值 +0000
P Difference to Greenwich time (GMT), with colon +00:00
T Timezone abbreviation. UTC
Z 时区偏移(秒)。 0
杂项
t Number of days in the current month. 30
c ISO 8601格式化日期,与Y-m-d"T"H:i:s+00:00等效。 2019-11-12T16:08:48+00:00
r RFC 5322 formatted date, equivalent to D, j M Y H:i:s +0000, with weekday name and month name not internationalized. Tue, 12 Nov 2019 16:08:48 +0000
非公历(格里历)历法
伊斯兰教历法
xmj 月内一日。 14
xmF 月的全名。 赖比尔·敖外鲁月
xmn 月索引。 3
xmY 该函数可嵌套。为了做到这一点,嵌套内层的#if函数需要按规则完整填写所有参数以符合#if的参数形式。一般至多可嵌套7层,该数也可能为wiki自身限制和内存限制而有变动。
$code1: 测试字符串 | 字符串非空时的值 | $code2: 测试字符串 | 字符串非空时的值 | 字符串为空(或只有空格)的值 $code3
1441
伊朗历法(Jalaly)
xit 月内日数。 30
xiz 年内一日。 236
xij 月内一日。 21
xiF 月的全名。 Aban
xin 月索引。 8
xiY 全年。 1398
xiy 2位数年。 98
希伯来历法
xjj 月内一日。 14
xjF 月的全名。 玛西班月
xjt 月内日数。 30
xjx Genitive form of the month name. Cheshvan
xjn 月数。 2
xjY 全年。 5780
泰历
xkY Full year in Thai solar calendar.
注意 注意: For years before 1941 the dates in Jan-Mar range are not calculated properly.
2562
民国历/主体历
xoY 全年。 108
日本历法
xtY 全年。 令和元
标记
xn Format the next numeric code as a raw ASCII number. 在印地语中,{{#time:H, xnH}}产生०६, 06。
xN Like xn, but as a toggled flag, which endures until the end of the string or until the next appearance of xN in the string.
xr Format the next number as a roman numeral. Only works for numbers up to 10,000
(upto 3,000 in pre MediaWiki 1.20).
{{#time:xrY}} → MMXIX
xh Format the next number as a Hebrew numeral. {{#time:xhY}} → ב'י"ט

This parser function takes a date and/or time (in the Gregorian calendar) and formats it according to the syntax given. A date/time object can be specified; the default is the value of the magic word {{CURRENTTIMESTAMP}} – that is, the time the page was last rendered into HTML.

{{#time: format string }}
{{#time: format string | date/time object }}
{{#time: format string | date/time object | language code }}
{{#time: format string | date/time object | language code | local }}

The list of accepted formatting codes is given in the table to the right. Any character in the formatting string that is not recognized is passed through unaltered; this applies also to blank spaces (the system does not need them for interpreting the codes). There are also two ways to escape characters within the formatting string:

  1. A backslash followed by a formatting character is interpreted as a single literal character
  2. Characters enclosed in double quotes are considered literal characters, and the quotes are removed.

In addition, the digraph xx is interpreted as a single literal "x".

{{#time: Y-m-d }}2019-11-12
{{#time: [[Y]] m d }}2019 11 12
{{#time: [[Y (year)]] }}2019 (19UTCpmTue, 12 Nov 2019 16:08:48 +0000)
{{#time: [[Y "(year)"]] }}2019 (year)
{{#time: i's" }}08'48"

The date/time object can be in any format accepted by PHP's strtotime() function. Both absolute (eg 20 December 2000) and relative (eg +20 hours) times are accepted.

{{#time: r|now}}Tue, 12 Nov 2019 16:08:48 +0000
{{#time: r|+2 hours}}Tue, 12 Nov 2019 18:08:48 +0000
{{#time: r|now + 2 hours}}Tue, 12 Nov 2019 18:08:48 +0000
{{#time: r|20 December 2000}}Wed, 20 Dec 2000 00:00:00 +0000
{{#time: r|December 20, 2000}}Wed, 20 Dec 2000 00:00:00 +0000
{{#time: r|2000-12-20}}Wed, 20 Dec 2000 00:00:00 +0000
{{#time: r|2000 December 20}}Error: Invalid time.

The language code in ISO 639-3 (?) allows the string to be displayed in the chosen language

{{#time:d F Y|1988-02-28|nl}}28 februari 1988
{{#time:l|now|uk}}вівторок
{{#time:d xg Y|20 June 2010|pl}}20 czerwca 2010

The local parameter specifies if the date/time object refers to the local timezone or to UTC.

This is a boolean parameters: its value is determined by casting the value of the argument (see the official PHP documentation for details on how string are cast to boolean values).

Please note that, if the variable $wgLocaltimezone is set to UTC, there is no difference in the output when local is set to true or false.

See the following examples for details:

{{#time: Y F d H:i:s|now|it|0}}2019 novembre 12 16:08:48
{{#time: Y F d H:i:s|now|it|1}}2019 novembre 12 16:08:48
{{#time: Y F d H:i:s|+2 hours||0}}2019 11月 12 18:08:48
{{#time: Y F d H:i:s|+2 hours||1}}2019 11月 12 18:08:48
{{#time:c|2019-05-16T17:05:43+02:00|it}}2019-05-16T15:05:43+00:00
{{#time:c|2019-05-16T17:05:43+02:00|it|0}}2019-05-16T15:05:43+00:00
{{#time:c|2019-05-16T17:05:43+02:00|it|true}}2019-05-16T15:05:43+00:00

If you've calculated a Unix timestamp, you may use it in date calculations by pre-pending an @ symbol.

{{#time: U | now }}1573574928
{{#time: r | @1573574928 }}Tue, 12 Nov 2019 16:08:48 +0000
警告 警告: 若不在數字格式的時間戳值的前面包含@前綴,結果通常會出錯或是未預期的數值:
{{#time: r | 1970-01-01 00:16:39 }}Thu, 01 Jan 1970 00:16:39 +0000
{{#time: U | 1970-01-01 00:16:39 }}999
{{#time: r | @999 }}Thu, 01 Jan 1970 00:16:39 +0000 (正確)
{{#time: r | 999 }}Error: Invalid time. (不支援的年分格式)
{{#time: r | 1970-01-01 00:16:40 }}Thu, 01 Jan 1970 00:16:40 +0000
{{#time: U | 1970-01-01 00:16:40 }}1000
{{#time: r | @1000 }}Thu, 01 Jan 1970 00:16:40 +0000 (正確)
{{#time: r | 1000 }}Wed, 12 Nov 1000 00:00:00 +0000 (被解讀為某年的目前月分和日期)
{{#time: r | 1970-01-01 02:46:39 }}Thu, 01 Jan 1970 02:46:39 +0000
{{#time: U | 1970-01-01 02:46:39 }}9999
{{#time: r | @9999 }}Thu, 01 Jan 1970 02:46:39 +0000 (正確)
{{#time: r | 9999 }}Fri, 12 Nov 9999 00:00:00 +0000 (被解讀為某年的目前月分和日期)
{{#time: r | 1970-01-01 02:46:40 }}Thu, 01 Jan 1970 02:46:40 +0000
{{#time: U | 1970-01-01 02:46:40 }}10000
{{#time: r | @10000 }}Thu, 01 Jan 1970 02:46:40 +0000 (正確)
{{#time: r | 10000 }}Error: Invalid time. (不支援的年分格式)
警告 警告: The range of acceptable input is 1 January 0111 → 31 December 9999. For the years 100 through 110 the output is inconsistent, Y and leap years are like the years 100-110, r, D, l and U are like interpreting these years as 2000-2010.

{{#time: d F Y | 29 Feb 0100 }}01 3月 0100
(正確,沒有閏年),但是
{{#time: r | 29 Feb 0100 }}Mon, 01 Mar 0100 00:00:00 +0000 (錯誤,100被解讀為2000,但那是閏年)
{{#time: d F Y | 15 April 10000 }}Error: Invalid time.
{{#time: r | 10000-4-15 }}Sat, 15 Apr 2000 10:00:00 +0000

Year numbers 0-99 are interpreted as 2000-2069 and 1970-1999, except when written in 4-digit format with leading zeros:
{{#time: d F Y | 1 Jan 6 }}01 1月 2006
{{#time: d F Y | 1 Jan 06 }}01 1月 2006
{{#time: d F Y | 1 Jan 006 }}01 1月 2006
{{#time: d F Y | 1 Jan 0006 }}01 1月 0006 (4-digit format)

The weekday is supplied for the years 100-110 and from 1753, for the years 111-1752 the r-output shows "Unknown" and the l-output "<>". As a consequence, the r-output is not accepted as input for these years.

Full or partial absolute dates can be specified; the function will "fill in" parts of the date that are not specified using the current values:

{{#time: Y | January 1 }}2019
警告 警告: The fill-in feature is not consistent; some parts are filled in using the current values, others are not:

{{#time: Y m d H:i:s | June }}2019 06 12 00:00:00 Gives the start of the day, but the current day of the month and the current year.

{{#time: Y m d H:i:s | 2003 }}2003 11 12 00:00:00 Gives the start of the day, but the current day of the year.

4位數字總是被解讀為年分,而不是小時和分鐘:[1]

{{#time: Y m d H:i:s | 1959 }}1959 11 12 00:00:00

6位數字盡量被解讀為小時、分鐘、秒,不然會被解讀為錯誤(而不是年分和月分):

{{#time: Y m d H:i:s | 195909 }}2019 11 12 19:59:09 輸入被視為時間而不是年+月代碼。
{{#time: Y m d H:i:s | 196009 }}Error: Invalid time. 雖然19:60:09不是有效時間,但196009不被解讀為1960年9月。

該函數執行一定數量的日期算數:

{{#time: d F Y | January 0 2008 }}31 12月 2007
{{#time: d F | January 32 }}Error: Invalid time.
{{#time: d F | February 29 2008 }}29 2月
{{#time: d F | February 29 2007 }}01 3月
{{#time:Y-F|now -1 months}}2019-10月

#time調用的格式字符串的總長度限制為6000個字符[2]

时区问题

There is a bug in this #time parser function (more specifically in PHP DateTime) that does not allow the passing-in of non-integers as relative time zone offsets. This issue does not apply when using an on-the-hour time zone, such as EDT. For example:

  • {{#time:g:i A | -4 hours }} ==> 12:08 PM

However, India is on a +5.5 hours time offset from UTC, and thus using its time zone will not normally allow the correct calculation of a relative time zone offset. Here's what happens:

  • {{#time:g:i A | +5.5 hours }} ==> 4:08 PM

To workaround this issue, simply convert the time into minutes or seconds, like this:

  • {{#time:g:i A | +330 minutes }} ==> 9:38 PM
  • {{#time:g:i A | +19800 seconds }} ==> 9:38 PM

(Tim Starling, the developer of this function, provided the exact syntax for this solution.)

#timel

This function is identical to {{#time: ... }}, when the local parameter is set to true, so it always uses the local time of the wiki (as set in $wgLocaltimezone ).

Syntax of the function is:

{{#timel: format string }}
{{#timel: format string | date/time object }}
{{#timel: format string | date/time object | language code }}
Please note that, if the variable $wgLocaltimezone is set to UTC, there is no difference in the output when local is set to true or false
Example of the use of #time and #timel parser functions from a server where the timezone is not UTC

For instance, see the following examples:

{{#time:c|now|it}}2019-11-12T16:08:48+00:00
{{#time:c|now|it|0}}2019-11-12T16:08:48+00:00
{{#time:c|now|it|1}}2019-11-12T16:08:48+00:00
{{#timel:c|now|it}}2019-11-12T16:08:48+00:00
警告 警告: Be aware that U for both time and timel will return the same number of seconds since 1970-01-01 00:00:00 UTC on Wikipedias with different timezones than UTC (formerly known as GMT)
U Unix时间。自1970年1月1日 00:00:00(GMT)以来的秒数。
Z 时区偏移(秒)。
{{#time: U}}1573574928
{{#timel: U}}1573574928
{{#time: Z}}0
{{#timel: Z}}0


#titleparts

This function separates a page title into segments based on slashes, then returns some of those segments as output.

{{#titleparts: pagename | number of segments to return | first segment to return }}

If the number of segments to return parameter is not specified, it defaults to "0", which returns all the segments from the first segment to return (included). If the first segment to return parameter is not specified or is "0", it defaults to "1":

{{#titleparts: Talk:Foo/bar/baz/quok }}Talk:Foo/bar/baz/quok
{{#titleparts: Talk:Foo/bar/baz/quok | 1 }}Talk:Foo See also {{ROOTPAGENAME }}.
{{#titleparts: Talk:Foo/bar/baz/quok | 2 }}Talk:Foo/bar
{{#titleparts: Talk:Foo/bar/baz/quok | 2 | 2 }}bar/baz
{{#titleparts: Talk:Foo/bar/baz/quok | | 2 }}bar/baz/quok
{{#titleparts: Talk:Foo/bar/baz/quok | | 5 }}

Negative values are accepted for both values. Negative values for the number of segments to return parameter effectively 'strips' segments from the end of the string. Negative values for the first segment to return translates to "start with this segment counting from the right":

{{#titleparts: Talk:Foo/bar/baz/quok | -1 }}Talk:Foo/bar/baz Strips one segment from the end of the string. See also {{BASEPAGENAME}}.
{{#titleparts: Talk:Foo/bar/baz/quok | -4 }} Strips all 4 segments from the end of the string
{{#titleparts: Talk:Foo/bar/baz/quok | -5 }} Strips 5 segments from the end of the string (more than exist)
{{#titleparts: Talk:Foo/bar/baz/quok | | -1 }} quok Returns last segment. See also {{SUBPAGENAME}}.
{{#titleparts: Talk:Foo/bar/baz/quok | -1 | 2 }} bar/baz Strips one segment from the end of the string, then returns the second segment and beyond
{{#titleparts: Talk:Foo/bar/baz/quok | -1 | -2 }} baz Start copying at the second last element; strip one segment from the end of the string

Before processing, the pagename parameter is HTML-decoded: if it contains some standard HTML character entities, they will be converted to plain characters (internally encoded with UTF-8, i.e. the same encoding as in the MediaWiki source page using this parser function).

For example, any occurence of &quot;, &#34;, or &#x22; in pagename will be replaced by ".
No other conversion from HTML to plain text is performed, so HTML tags are left intact at this initial step even if they are invalid in page titles.
Some magic keywords or parser functions of MediaWiki (such as {{PAGENAME }} and similar) are known to return strings that are needlessly HTML-encoded, even if their own input parameter was not HTML-encoded:

The titleparts parser function can then be used as a workaround, to convert these returned strings so that they can be processed correctly by some other parser functions also taking a page name in parameter (such as {{PAGESINCAT: }} but which are still not working properly with HTML-encoded input strings.

For example, if the current page is Category:Côte-d'Or, then:

  • {{#ifeq: {{FULLPAGENAME}} | Category:Côte-d'Or | 1 | 0 }}, and {{#ifeq: {{FULLPAGENAME}} | Category:Côte-d&apos;Or | 1 | 0 }} are both returning 1; (the #ifeq parser function does perform the HTML-decoding of its input parameters).
  • {{#switch: {{FULLPAGENAME}} | Category:Côte-d'Or = 1 | #default = 0 }}, and {{#switch: {{FULLPAGENAME}} | Category:Côte-d&apos;Or = 1 | #default = 0 }} are both returning 1; (the #switch parser function does perform the HTML-decoding of its input parameters).
  • {{#ifexist: {{FULLPAGENAME}} | 1 | 0 }}, {{#ifexist: Category:Côte-d'Or | 1 | 0 }}, or even {{#ifexist: Category:Côte-d&apos;Or | 1 | 0 }} will all return 1 if that category page exists (the #ifexist parser function does perform the HTML-decoding of its input parameters);
  • {{PAGESINCAT: Côte-d'Or }} will return a non-zero number, if that category contains pages or subcategories, but:
  • {{PAGESINCAT: {{CURRENTPAGENAME}} }}, may still unconditionally return 0, just like:
  • {{PAGESINCAT: {{PAGENAME|Category:Côte-d'Or}} }}
  • {{PAGESINCAT: {{PAGENAME|Category:Côte-d&apos;Or}} }}

The reason of this unexpected behavior is that, with the current versions of MediaWiki, there are two caveats:

  • {{FULLPAGENAME}}, or even {{FULLPAGENAME|Côte-d'Or}} may return the actually HTML-encoded string Category:Côte-d&apos;Or and not the expected Category:Côte-d'Or, and that:
  • {{PAGESINCAT: Côte-d&apos;Or }} unconditionally returns 0 (the PAGESINCAT magic keyword does not perform any HTML-decoding of its input parameter).

The simple workaround using titleparts (which will continue to work if the two caveats are fixed in a later version of MediaWiki) is:

  • {{PAGESINCAT: {{#titleparts: {{CURRENTPAGENAME}} }} }}
  • {{PAGESINCAT: {{#titleparts: {{PAGENAME|Category:Côte-d'Or}} }} }}
  • {{PAGESINCAT: {{#titleparts: {{PAGENAME|Category:Côte-d&apos;Or}} }} }}, that all return the actual number of pages in the same category.

Then the decoded pagename is canonicalized into a standard page title supported by MediaWiki, as much as possible:

  1. All underscores are automatically replaced with spaces:
    {{#titleparts: Talk:Foo/bah_boo|1|2}}bah boo Not bah_boo, despite the underscore in the original.
  2. The string is split a maximum of 25 times; further slashes are ignored and the 25th element will contain the rest of the string. The string is also limited to 255 characters, as it is treated as a page title:
    {{#titleparts: a/b/c/d/e/f/g/h/i/j/k/l/m/n/o/p/q/r/s/t/u/v/w/x/y/z/aa/bb/cc/dd/ee | 1 | 25 }}y/z/aa/bb/cc/dd/ee
    If for whatever reason you needed to push this function to its limit, although very unlikely, it is possible to bypass the 25 split limit by nesting function calls:
    {{#titleparts: {{#titleparts: a/b/c/d/e/f/g/h/i/j/k/l/m/n/o/p/q/r/s/t/u/v/w/x/y/z/aa/bb/cc/dd/ee| 1 | 25 }} | 1 | 2}}z
  3. Finally the first substring is capitalized according to the capitalization settings of the local wiki (if that substring also starts by a local namespace name, that namespace name is also normalized).
    {{#titleparts: talk:a/b/c }}Talk:A/b/c
警告 警告: You can use #titleparts as a small "string parser and converter", but consider that it returns the first substring capitalized:
{{#titleparts: one/two/three/four|1|1 }}One
{{#titleparts: one/two/three/four|1|2 }}two

If lower case is needed, use lc: function to control output:

{{lc: {{#titleparts: one/two/three/four|1|1 }} }}one

You can prepend a 'dummy' slash at the beginning of the string to get the correct first substring capitalization (uppercase or lowercase). Use 2 instead of 1 for first segment to return:

{{#titleparts: /one/two/three/four|1|2 }}one
{{#titleparts: /One/two/three/four|1|2 }}One
警告 警告: Certain characters that are illegal in a page title will cause #titleparts to not parse the string:
{{#titleparts: {one/two} | 1 | 1 }}{one/two}. Does not produce the expected: {one
{{#titleparts: [[page]]/123 | 1 | 2 }}page/123. Does not work because brackets are illegal in page titles and this parser function does not process links embedded in its input pagename parameter, even when they use the MediaWiki syntax, or any other HTML or MediaWiki tags.
{{#titleparts: red/#00FF00/blue | 1 | 3 }} → "". Does not work because "#" is also illegal in page titles.
警告 警告: If any part of the title is just "." or "..", #titleparts will not parse the string:
{{#titleparts: one/./three | 1 | 1 }}one/./three. The whole string is returned. It does not produce the expected: one
警告 警告: This function does not degrade gracefully if the input exceeds 255 characters. If the input string is 256 characters long or more, the whole string is returned.

一般帮助

替代

Parser functions can be substituted by prefixing the hash character with subst::

{{subst:#ifexist: Help:Extension:ParserFunctions/zh | [[Help:Extension:ParserFunctions/zh]] | Help:Extension:ParserFunctions/zh }} → the code [[Help:Extension:ParserFunctions/zh]] will be inserted in the wikitext since the page Help:Extension:ParserFunctions/zh exists.
警告 警告: The results of substituted parser functions are undefined if the expressions contain unsubstituted volatile code such as variables or other parser functions. For consistent results, all the volatile code in the expression to be evaluated must be substituted. See Help:Substitution.

Substitution does not work within <ref></ref> , you can use {{subst:#tag:ref|}} for this purpose.

重定向

Especially {{#time:…|now-…}} could be handy in redirects to pages including dates, but this does not work.

在表格中逸出的管道符

Parser functions will mangle wikitable syntax and pipe characters (|), treating all the raw pipe characters as parameter dividers. To avoid this, most wikis used a template Template:! with its contents only a raw pipe character (|), since MW 1.24 a {{!}} magic word replaced this kludge. This 'hides' the pipe from the MediaWiki parser, ensuring that it is not considered until after all the templates and variables on a page have been expanded. It will then be interpreted as a table row or column separator. Alternatively, raw HTML table syntax can be used, although this is less intuitive and more error-prone.

You can also escape the pipe character for display as a plain, uninterpreted character using an HTML entity: &#124; .

说明 您输入的 您输出的
Escaping pipe character as table row/column separator
{{!}}
|
Escaping pipe character as a plain character
&#124;
|

剥离空格

Whitespace, including newlines, tabs, and spaces, is stripped from the beginning and end of all the parameters of these parser functions. If this is not desirable, comparison of strings can be done after putting them in quotation marks.

{{#ifeq: foo           |           foo | equal | not equal }}equal
{{#ifeq: "foo          " | "          foo" | equal | not equal }}not equal

To prevent the trimming of then and else parts, see m:Template:If. Some people achieve this by using <nowiki> </nowiki> instead of spaces.

foo{{#if:|| bar }}foofoobarfoo
foo{{#if:||<nowiki /> bar <nowiki />}}foofoo bar foo

However, this method can be used to render a single whitespace character only, since the parser squeezes multiple whitespace characters in a row into one.

<span style="white-space: pre;">foo{{#if:||<nowiki></nowiki>      bar      <nowiki></nowiki>}}foo</span>
foo bar foo

In this example, the white-space: pre style is used to force the whitespace to be preserved by the browser, but even with it the spaces are not shown. This happens because the spaces are stripped by the software, before being sent to the browser.

It is possible to workaround this behavior replacing whitespaces with &#32; (breakable space) or &nbsp; (non-breakable space), since they are not modified by the software:

<span style="white-space: pre;">foo{{#if:||&#32;&#32;&#32;bar&#32;&#32;&#32;}}foo</span>foo bar foo
foo{{#if:||&nbsp;&nbsp;&nbsp;bar&nbsp;&nbsp;&nbsp;}}foofoo   bar   foo

参见

参考资料

  1. 在2011年r86805之前,情況並非如此。
  2. phabricator.wikimedia.org的ExtParserFunctions.php