# Help:Extension:ParserFunctions

This page is a translated version of the page Help:Extension:ParserFunctions and the translation is 50% complete.

Other languages:
العربية • ‎български • ‎বাংলা • ‎català • ‎Deutsch • ‎Zazaki • ‎English • ‎español • ‎فارسی • ‎suomi • ‎français • ‎galego • ‎italiano • ‎日本語 • ‎한국어 • ‎Lëtzebuergesch • ‎मराठी • ‎polski • ‎پښتو • ‎português • ‎português do Brasil • ‎русский • ‎svenska • ‎தமிழ் • ‎ไทย • ‎Türkçe • ‎Tiếng Việt • ‎中文

The ParserFunctions extension provides eleven additional parser functions to supplement the "magic words", which are already present in MediaWiki. All the parser functions provided by this extension take the form:

`{{#関数名: 引数 1 | 引数 2 | 引数 3 ... }}`

## #expr

グループ化 (括弧) `( )`

`* / div mod`
`+ -`

`and`
`or`

この関数は、数式を計算してその計算結果を返します。

`{{#expr: 数式 }}`

The available operators are listed to the right, in order of precedence. See Help:Calculation for more details of the function of each operator. The accuracy and format of the result returned will vary depending on the operating system of the server running the wiki, and the number format of the site language.

When evaluating using boolean algebra, zero evaluates to `false` and any nonzero value, positive or negative, evaluates to `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

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

The order of addition and subtraction operands before or after a number is meaningful, and may be treated as a positive or negative value instead of as an operand with an erroneous input:

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

Note, if using the output of magic words, you must raw format them in order to remove commas, and translate the numerals. For example {{NUMBEROFUSERS}} results in 2,757,028, where we want 2757028, which can be obtained using `{{formatnum:{{NUMBEROFUSERS}}|R}}`. This is especially important in some languages, where numerals are translated. For example, in Bengali, {{NUMBEROFUSERS}} produces ৩০,০৬১.

`{{#expr:{{NUMBEROFUSERS}}+100}}`Expression error: Unrecognized punctuation character ",".
`{{#expr:{{formatnum:{{NUMBEROFUSERS}}|R}}+100}}`2757128
 警告: 演算子 `mod` は、第 2 引数の値によっては誤った結果を返す場合があります: `{{#expr: 123 mod (2^64-1)}}` → Division by zero. (空の文字列を返しますが、123 を返すべきです)
If you want to do calculations based on dates (ex. test if current date and time is after some other date and time), first convert the time to number of seconds after January 1, 1970 using {{#time: xNU }}, then you can simply add and subtract dates as numbers.

### 丸め

Rounds off the number on the left to a multiple of 1/10 raised to a power, with the exponent equal to the truncated value of the number given on the right.

`{{#expr: 1/3 round 5 }}` 0.33333 最終桁が 5 未満であるため、切り上げはされません
`{{#expr: 1/6 round 5 }}` 0.16667 最終桁が 5 以上であるため、切り上げされます
`{{#expr: 8.99999/9 round 5 }}` 1 Again, the result is rounded up on the last digit, which results in additional rounding
`{{#expr: 1234.5678 round -2 }}` 1200 Rounded to nearest 100 because negative values round to the left of the decimal point
`{{#expr: 1234.5678 round 2 }}` 1234.57 Rounded to nearest 100th since positive values round to the right of the decimal point
`{{#expr: 1234.5678 round 2.3 }}` 1234.57 round の右の値の小数点以下は、丸めの結果には影響しません
`{{#expr: trunc 1234.5678 }}` 1234 小数点以下の切り捨て

`{{#expr: 1/3 round 0 }}` 0 最も近い整数への切り捨てで 0
`{{#expr: 1/2 round 0 }}` 1 最も近い整数への切り上げで 1
`{{#expr: 3/4 round 0 }}` 1 最も近い整数への切り上げで 1
`{{#expr: -1/3 round 0 }}` 0 最も近い整数への切り上げで 0
`{{#expr: -1/2 round 0 }}` -1 最も近い整数への切り捨てで -1
`{{#expr: -3/4 round 0 }}` -1 最も近い整数への切り捨てで -1
ceillfloor による切り上げや切り捨て
`{{#expr: ceil(1/3) }}` 1 Up to the next larger integer, which is one
`{{#expr: floor(1/3) }}` 0 Down to the next smaller integer, which is zero
`{{#expr: ceil(-1/3) }}` 0 Up to the next larger integer, which is zero
`{{#expr: floor(-1/3) }}` -1 Down to the next smaller integer, which is negative one
`{{#expr: ceil 1/3 }}` 0.33333333333333 1 が既に整数であるため、丸めは行なわれない
 警告: あなたの予想とは異なり、ceil(1/3) ではなく (ceil 1)/3 と解釈される

## #if

この関数はテスト文字列を評価し、それが空であるかどうかを判定します。空白のみを含むテスト文字列は空であると見做されます。

`{{#if: テスト文字列 | テスト文字列が空ではない場合の値 | テスト文字列が空 (または空白のみ) の場合の値 }}`
`{{#if: パラメーター 1 | パラメーター 2 | パラメーター 3 }}`

この関数は、初めに最初の引数が空でないか確かめます。最初の引数が空でない場合、2 番目の引数を表示します。最初の引数が空または空白文字（空白、改行等）のみを含む場合、3 番目の引数を表示します。

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

テスト文字列は常に純粋なテキストとして評価されます。そのため数式は評価されません:

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

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

The function may be nested. To do so, nest the inner #if function in its full form in place of a parameter of the enclosing #if function. A depth of at least seven levels of nesting is possible, although that may depend on the wiki or a memory limit.

• Example: {{#if: test string | value if test string is not empty | {{#if: test string | value if test string is not empty | value if test string is empty (or only white space) }} }}
1. if 文において引数をテスト文字列として用いることも可能です。変数名の後には |（パイプ記号）を追加する必要があります。
• Example `{{#if:{{{1|}}}|You entered text in variable 1|There is no text in variable 1}}`

このパーサー関数の他の例については Help:Parser functions in templates/ja を参照してください。

## #ifeq

This parser function compares two strings and determines whether they are identical.

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

If both strings are valid numerical values, the strings are compared numerically:

`{{#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

Otherwise the comparison is made as text; this comparison is case sensitive:

`{{#ifeq: foo | bar | equal | not equal}}`not equal
`{{#ifeq: foo | Foo | equal | not equal}}`not equal
`{{#ifeq: "01" | "1" | equal | not equal}}`not equal  (compare to similar example above, without the quotes)
`{{#ifeq: 10^3 | 1000 | equal | not equal}}`not equal  (compare to similar example above, with `#expr`)

As a practical example, given an existing template `Template:Size` used to set standard short and long sizes, defined as:

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

the following ensue:

`{{size|short}}`20
`{{size|20}}`40
`{{size}}`40
 警告: Numerical comparisons with `#ifeq` and `#switch` are not equivalent to comparisons with `#expr`: `{{#ifeq: 12345678901234567 | 12345678901234568 | equal | not equal}}` → not equal `{{#switch: 12345678901234567 | 12345678901234568 = equal | not equal}}` → not equal because PHP compares two numbers of type integer, whereas: `{{#ifexpr: 12345678901234567 = 12345678901234568 | equal | not equal}}` → equal because MediaWiki converts literal numbers in expressions to type float, which, for large integers like these, involves rounding.
 警告: Tags and parser functions inside parser tags (such as ``) are temporarily replaced by a unique code. This affects comparisons: `{{#ifeq: foo | foo | equal | not equal}}` → not equal `{{#ifeq: [itex]foo[/itex] | [itex]foo[/itex] | equal | not equal}}` → not equal `{{#ifeq: {{#tag:math|foo}} | {{#tag:math|foo}} | equal | not equal}}` → not equal `{{#ifeq: [[foo]] | [[foo]] | equal | not equal}}` → equal If the strings to be compared are given as equal calls to the same template containing such tags, then the condition is true, but in the case of two templates with identical content containing such tags it is false.
 警告: Be careful when comparing against the current page title using the page name magic words. These magic words convert special characters into numeric HTML entities. This may result in misleading results. For example, if you are on a page titled "L'Aquila"... `{{#ifeq: L'Aquila | {{FULLPAGENAME}} | equal | not equal}}` → not equal[dubious ] To work around this, apply the magic word to both parameters: `{{#ifeq: {{FULLPAGENAME: L'Aquila}} | {{FULLPAGENAME}} | equal | not equal}}` → equal

## #iferror

この関数は入力文字列を取り、2つの結果から1つを返します。この関数は、 `#expr``#time``#rel2abs` のような他のパーサー関数、ループや再帰のようなテンプレートエラー、その他の「フェースソフト」なパーサーエラーによって作られた `class="error"` を含む HTML オブジェクトが入力文字列に含まれていた場合に `true` と評価します。

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

`{{#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

This function evaluates a mathematical expression and returns one of two strings depending on the boolean value of the result:

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

The `expression` input is evaluated exactly as for `#expr` above, with the same operators being available. The output is then evaluated as a boolean expression.

An empty input expression evaluates to `false`:

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

As mentioned above, zero evaluates to `false` and any nonzero value evaluates to `true`, so this function is equivalent to one using `#ifeq` and `#expr` only:

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

except for an empty or wrong input expression (an error message is treated as an empty string; it is not equal to zero, so we get `value if true`).

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

comparing

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

Either or both the return values may be omitted; no output is given when the appropriate branch is left empty:

`{{#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

This function takes an input string, interprets it as a page title, and returns one of two values depending on whether or not the page exists on the local wiki.

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

The function evaluates to `true` if the page exists, whether it contains content, is visibly blank (contains meta-data such as category links or magic words, but no visible content), is blank, or is a redirect. Only pages that are redlinked evaluate to `false`, including if the page used to exist but has been deleted.

`{{#ifexist: Help:Extension:ParserFunctions/ja | exists | doesn't exist }}`exists
`{{#ifexist: XXHelp:Extension:ParserFunctions/jaXX | exists | doesn't exist }}`doesn't exist

The function evaluates to `true` for system messages that have been customized, and for special pages that are defined by the software.

`{{#ifexist: Special:Watchlist | exists | doesn't exist }}`exists
`{{#ifexist: Special:CheckUser | exists | doesn't exist }}`exists (because the CheckUser extension is installed on this wiki)
`{{#ifexist: MediaWiki:Copyright | exists | doesn't exist }}`exists (because MediaWiki:Copyright has been customized)

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/ja), Special:WhatLinksHere/Foo will list Help:Extension:ParserFunctions/ja.

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:` does not work with interwiki links.

### ifexistの制限

`#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.

## #rel2abs

この関数は相対ファイル パスを絶対ファイル パスに変換します。

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

`path` 部分では、以下の構文を使用できます:

• `.` → 現在の階層
• `..` → 「1 つ上の階層に移動」
• `/foo` → 「下位ディレクトリ /foo に 1 階層移動」

`base path` を指定していない場合は、ページの完全なページ名が代わりに使用されます:

`{{#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
}}
```

`{{#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.

### 既定

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 }}`

### Grouping results

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`

### 比較の挙動

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: Numerical comparisons with `#switch` and `#ifeq` are not equivalent with comparisons in expressions (see also above): `{{#switch: 12345678901234567 | 12345678901234568 = A | B}}` → B `{{#ifexpr: 12345678901234567 = 12345678901234568 | A | B}}` → A

### 生の等号

"Case" strings cannot contain raw equals signs. To work around this, create a template {{=}} containing a single equals sign: `=`.

`{{#switch: 1=2`
` | 1=2 = raw`
` | 1<nowiki>=</nowiki>2 = nowiki`
` | 1&#61;2 = html`
` | 1{{=}}2 = template`
` | default`
`}}`html

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

### #ifeqの置換

`#switch` can be used to reduce expansion depth.

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

は、以下と同等です

• `{{#ifeq:{{{1}}}|condition1 |branch1 |{{#ifeq:{{{1}}}|condition2 |branch2 |{{#ifeq:{{{1}}}|condition3 |branch3 |branch4}}}}}}`
(i.e. `{{#ifeq:{{{1}}}|condition1 |<!--then-->branch1 |<!--else-->{{#ifeq:{{{1}}}|condition2 |<!--then-->branch2 |<!--else-->{{#ifeq:{{{1}}}|condition3 |<!--then-->branch3 |<!--else-->branch4}}}}}}`)

## #time

コード 説明 現在の出力
(更新するにはこのページのキャッシュを破棄してください)
`Y` 4 桁の年 2015
`y` 2 桁の年 15
`L` 閏年の場合 1、平年の場合 0 0
`o` ¹ 指定した週の ISO-8601 年 ² 2015 ³

¹ PHP 5.1.0 以降および rev:45208 が必要
² This has the same value as Y, except that if the ISO week number (W) belongs to the previous or next year, that year is used instead.
³ Will output literal o if ¹ not fulfilled.

`n` 月番号、ゼロ補完なし 7
`m` 月番号、ゼロ補完あり 07
`M` サイトの言語での月名の省略形 7月
`F` サイトの言語での完全な月名 7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. ポーランド語の例:

(主格)

`{{#time:d F Y|20 June 2010|pl}}` → 20 czerwiec 2010

(所有格)

`{{#time:d xg Y|20 June 2010|pl}}` → 20 czerwca 2010

`W` ISO 8601 週番号、ゼロ補完あり 31

`j` 日、ゼロ補完なし 28
`d` 日、ゼロ補完あり 28
`z` Day of the year (January 1 = 0).
Note: To get the ISO day of the year add 1.
208
`D` An abbreviation for the day of the week. Rarely internationalized.
`l` The full weekday name. Rarely internationalized. 火曜日
`N` ISO 8601 曜日番号 (月曜日 = 1、日曜日 = 7) 2
`w` 曜日番号 (日曜日 = 0、土曜日 = 6) 2
`a` 午前 (00:00:00 ～ 11:59:59) の場合「am」、午後 (12:00:00 ～ 23:59:59) の場合「pm」 pm
`A` 上の `a` の大文字版 PM
`g` 時、12 時間制、ゼロ補完なし 10
`h` 時、12 時間制、ゼロ補完あり 10
`G` 時、24 時間制、ゼロ補完なし 22
`H` 時、24 時間制、ゼロ補完あり 22

`i` 分、ゼロ補完あり 12
`s` 秒、ゼロ補完あり 52
`U` 1970年1月1日 00:00:00 GMT からの経過秒数 1438121572
タイムゾーン (1.22wmf2 時点)
`e` タイムゾーン識別子 UTC
`I` 日時がサマータイムかどうか 0
`O` グリニッジ標準時 (GMT) との差 +0000
`P` グリニッジ標準時 (GMT) との差、コロン付き +00:00
`T` タイムゾーンの省略形 UTC
`Z` タイムゾーンのオフセット (秒) 0
その他
`t` 31
`c` ISO 8601 形式の日付、`Y-m-d"T"H:i:s+00:00` と同等 2015-07-28T22:12:52+00:00
`r` RFC 5322 形式の日時、`D, j M Y H:i:s +0000` と同等。曜日と月名は国際化されない。 Tue, 28 Jul 2015 22:12:52 +0000
グレゴリオ暦以外
イスラム暦
`xmj` 11
`xmF` 完全な月名 シャウワール
`xmn` 月番号 10
`xmY` 完全な年 1436
イラン暦 (Jalaly)
`xij` 6
`xiF` 完全な月名 モルダード
`xin` 月番号 5
`xiY` 完全な年 1394
`xiy` 2 桁の年 94
ユダヤ暦
`xjj` 12
`xjF` 完全な月名 アブ
`xjt` 月の日数 30
`xjx` 月名の所有格 アブ
`xjn` 月番号 11
`xjY` 完全な年 5775
タイ太陽暦
`xkY` タイ太陽暦の完全な年。
: 1941 年より前の 1 月から 3 月の範囲は、適切には計算できません。
2558

`xoY` 完全な年 104

`xtY` 完全な年 平成27
フラグ
`xn` 後続の数値をそのまま半角英数字に整形する。 ヒンディー語では、`{{#time:H, xnH}}` は ०६, 06 を返す。
`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` 後続の数値をローマ数字に整形する。10,000 までの数値で動作する。
(MediaWiki 1.20 より前では 3,000 まで)
`{{#time:xrY}}` → MMXV
`xh` 後続の数値をヘブライ数字に整形する。 `{{#time:xhY}}` → ב'ט"ו

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 }}`

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 }}`2015-07-28
`{{#time: [[Y]] m d }}`2015 07 28
`{{#time: [[Y (year)]] }}`2015 (15UTCpmTue, 28 Jul 2015 22:12:52 +0000)
`{{#time: [[Y "(year)"]] }}`2015 (year)
`{{#time: i's" }}`12'52"

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, 28 Jul 2015 22:12:52 +0000
`{{#time: r|+2 hours}}`Wed, 29 Jul 2015 00:12:52 +0000
`{{#time: r|now + 2 hours}}`Wed, 29 Jul 2015 00:12:52 +0000

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

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

`{{#time: U | now }}`1438121572
`{{#time: r|@1438121572}}`Tue, 28 Jul 2015 22:12:52 +0000

 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 (correct, no leap year), but `{{#time: r | 29 Feb 0100 }}` → Mon, 01 Mar 0100 00:00:00 +0000 (wrong, even if 100 is interpreted as 2000, because that is a leap year) `{{#time: d F Y | 15 April 10000 }}` → Error: Invalid time. `{{#time: r | 10000-4-15 }}` → Sat, 15 Apr 2000 10:00:00 +0000 先頭に 0 を付けて 4 桁の書式にした場合を除いて、年の値 0-99 は 2000-2069 年および 1970-1999 年と解釈されます: `{{#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 }}`2015
 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 }}` → 2015 06 28 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 07 28 00:00:00 Gives the start of the day, but the current day of the year.

With MediaWiki r86805 - Code Review, a four-digit number is always interpreted as a year, never as hours and minutes:
`{{#time: Y m d H:i:s | 1959 }}`1959 07 28 00:00:00

A six-digit number is interpreted as hours, minutes and seconds if possible, but otherwise as an error (not, for instance, a year and month):
`{{#time: Y m d H:i:s | 195909 }}`2015 07 28 19:59:09 Input is treated as a time rather than a year+month code.
`{{#time: Y m d H:i:s | 196009 }}`Error: Invalid time. 19:60:09 は有効な時刻ではないにもかかわらず、196009 は1960年9月とは解釈されません。

The function performs a certain amount of date mathematics:

`{{#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}}`2015-6月

The total length of the format strings of the calls of `#time` is limited to 6000 characters [1].

#### タイムゾーンの問題点

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 }} ==> 6:12 PM

However, Venezuela is on a -4.5 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 | -4.5 hours }} ==> 7:12 AM

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

• {{#time:g:i A | -270 minutes }} ==> 5:42 PM
• {{#time:g:i A | -16200 seconds }} ==> 5:42 PM

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

## #timel

This function is identical to `{{#time: ... }}`, except that it uses the local time of the wiki (as set in \$wgLocaltimezone) when no date is given.

`{{#time: Y-m-d }}`2015-07-28
`{{#timel: Y-m-d }}`2015-07-28
`{{#time: Y F d H:i:s}}`2015 7月 28 22:12:52
`{{#timel: Y F d H:i:s}}`2015 7月 28 22:12:52

## #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

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. The list of substrings is not cleaned up hierarchically (unlike with the `{{PAGENAME}}` magic keyword, the `./` and `../` relative path segments are left intact by #titleparts):
`{{#titleparts: ../c/d }}`../c/d
`{{PAGENAME: ../c/d }}` → '
4. 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: 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: Certain characters that are illegal in a page title will cause #titleparts to not parse the string: `{{#titleparts: {one/two} | 1 | 1 }}` → {one/two}. 以下の値にはなりません: {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: This function does not degrade gracefully if your input exceeds 255 characters. If the inputted string is 256 characters long or more, this function will simply toss the string back at you.

## General points

### subst展開

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

`{{subst:#ifexist: Help:Extension:ParserFunctions/ja | [[Help:Extension:ParserFunctions/ja]] | Help:Extension:ParserFunctions/ja }}` → the code `[[Help:Extension:ParserFunctions/ja]]` will be inserted in the wikitext since the page Help:Extension:ParserFunctions/ja exists.
 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.

### リダイレクト

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

### 表内でのパイプ記号のエスケープ

Parser functions will mangle wikitable syntax and pipe characters (`|`), treating all the raw pipe characters as parameter dividers. To avoid this, most wikis create the template Template:! with its contents only a raw pipe character (`|`) or use the `{{!}}` magic word which is available since MW 1.24. 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.

パイプ記号は、プレーンな (解釈されない) 文字としてエスケープすることもできます。これには HTML エンティティ `&#124;` を使用します。

パイプ記号を表の行/列区切りとしてエスケープ
```{{!}}
```
|
パイプ記号をプレーンな文字としてエスケープ
```&#124;
```
|

### 空白類の除去

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

then 部分および else 部分での空白の除去を回避する方法は、m:Template:If を参照してください。空白の代わりに <nowiki> </nowiki> を使用することもできます。

`foo{{#if:|| bar }}foo`foobarfoo
`foo{{#if:||<nowiki></nowiki> bar <nowiki></nowiki>}}foo`foo bar foo

But this method allows the rendering of only one whitespace, since the parser reduce multiple spaces to a single 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;}}foo`foo   bar   foo