Help:Extension:ParserFunctions

From MediaWiki.org
Jump to: navigation, search

Other languages:
български • ‎বাংলা • ‎Deutsch • ‎Zazaki • ‎English • ‎British English • ‎español • ‎français • ‎日本語 • ‎Lëtzebuergesch • ‎मराठी • ‎polski • ‎پښتو • ‎português • ‎português do Brasil • ‎русский • ‎svenska • ‎தமிழ் • ‎ไทย • ‎Türkçe • ‎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:

{{#functionname: argument 1 | argument 2 | argument 3 ... }}

#expr[edit | edit source]

Type Operators
Grouping (parentheses) ( )
Numbers 1234.5   e (2.718)   pi (3.142)
binary operator e   unary +,-
Unary not ceil trunc floor abs exp ln sin cos tan acos asin atan
Binary ^
* / div mod
+ -
Round round
Logic = != <> > < >= <=
and
or

This function evaluates a mathematical expression and returns the calculated value.

{{#expr: expression }}

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

An empty input expression returns an empty string. Invalid expressions return one of several error messages, which can be caught using the #iferror function:

{{#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 {{NUMBEROFUSERS}}, where we want {{NUMBEROFUSERS}} which can be obtained by {{NUMBEROFUSERS}}. 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}}1791563


Warning Warning: The operator mod gives wrong results for some values of the second argument:
{{#expr: 123 mod (2^64-1)}}Division by zero. (produces an empty string; should be 123)
If you want to do calculations based on dates (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: U }}, then you can simply add and subtract dates.

Rounding[edit | edit source]

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.

To round up or down use unary ceil or floor respectively.

Test case Result Method of rounding
{{#expr: 1/3 round 5 }} 0.33333 Final digit is < 5, so no apparent rounding occurs
{{#expr: 1/6 round 5 }} 0.16667 Final digit is >= 5, so it is rounded up
{{#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 Decimals in the rounding index make no difference in the rounded result
{{#expr: trunc 1234.5678 }} 1234 Decimal portion truncated (chopped off)
Rounding to the nearest integer
{{#expr: 1/3 round 0 }} 0 Down to the nearest integer, which is zero
{{#expr: 1/2 round 0 }} 1 Up to the nearest integer, which is one
{{#expr: 3/4 round 0 }} 1 Up to the nearest integer, which is one
{{#expr: -1/3 round 0 }} -0 Up to the nearest integer, which is zero
Warning Warning: The negative sign on zero could cause equality mismatches
{{#expr: -1/2 round 0 }} -1 Down to the nearest integer, which is negative one
{{#expr: -3/4 round 0 }} -1 Down to the nearest integer, which is negative one
Rounding up or down with ceil and floor
{{#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
Warning Warning: The negative sign on zero could cause equality mismatches
{{#expr: floor(-1/3) }} -1 Down to the next smaller integer, which is negative one
{{#expr: ceil 1/3 }} 0.33333333333333 Not rounded since 1 already is an integer
Warning Warning: Interpreted as (ceil 1)/3, not ceil(1/3) as you might expect

#if[edit | edit source]

This function evaluates a test string and determines whether or not it is empty. A test string containing only white space is considered to be empty.

{{#if: test string | value if test string is not empty | value if test string is empty (or only white space) }}
{{#if: first parameter | second parameter | third parameter }}

This function first tests whether the first parameter is not empty. If the first parameter is not empty the function displays the second argument. If the first parameter is empty or contains only whitespace characters (spaces, newlines, etc.) it displays the third argument.

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


| yes | no}}
no

The test string is always interpreted as pure text, so mathematical expressions are not evaluated:

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

Either or both the return values may be omitted:

{{#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 the third 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) }} }}

See Help:Parser functions in templates for more examples of this parser function.

#ifeq[edit | edit source]

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
Warning Warning: 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.

Warning Warning: Tags and parser functions inside parser tags (such as <nowiki>) are temporarily replaced by a unique code. This affects comparisons:
{{#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

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.

Warning Warning:

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

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

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

#iferror[edit | edit source]

This function takes an input string and returns one of two results; the function evaluates to true if the input string contains an HTML object with class="error", as generated by other parser functions such as #expr, #time and #rel2abs, template errors such as loops and recursions, and other "failsoft" parser errors.

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

One or both of the return strings can be omitted. If the correct string is omitted, the test string is returned if it is not erroneous. If the error string is also omitted, an empty string is returned on an error:

{{#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: <strong class="error">a</strong> | error | correct }}error

#ifexpr[edit | edit source]

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[edit | edit source]

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 | exists | doesn't exist }}exists
{{#ifexist: XXHelp:Extension:ParserFunctionsXX | 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), Special:WhatLinksHere/Foo will list Help:Extension:ParserFunctions.

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 limits[edit | edit source]

#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[edit | edit source]

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[edit | edit source]

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
}}

Examples:

{{#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 allows an editor [to add information in one template and this information will be visible in several other templates which all have different formatting.][clarification needed][examples needed]

Default[edit | edit source]

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[edit | edit source]

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 = result2
 | case5 = result3
 | case6 
 | case7 = result4
 | #default = default result
}}

Here cases 2, 3 and 4 all return result2; cases 6 and 7 both return result4

Comparison behavior[edit | edit source]

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
Warning Warning: 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

Raw equal signs[edit | edit source]

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

Example:

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

Note 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.

Replacing #ifeq[edit | edit source]

#switch can be used to reduce expansion depth.

For example:

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

is equivalent to

  • {{#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[edit | edit source]

Code Description Current output
(Purge this page's cache to update)
Year
Y 4-digit year. 2014
y 2-digit year. 14
L 1 if it's a leap year, 0 if not. 0
o ¹ ISO-8601 year of the specified week. ² 2014 ³

¹ Requires PHP 5.1.0 and newer and rev:45208.
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.

Month
n Month index, not zero-padded. 9
m Month index, zero-padded. 09
M An abbreviation of the month name, in the site language. Sep
F The full month name in the site language. September
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. For Polish:

(nominative)

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

(genitive)

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

Week
W ISO 8601 week number, zero-padded. 38
Day
j Day of the month, not zero-padded. 21
d Day of the month, zero-padded. 21
z Day of the year (January 1 = 0).
Note Note: To get the ISO day of the year add 1.
263
D An abbreviation for the day of the week. Rarely internationalized. Sun
l The full weekday name. Rarely internationalized. Sunday
N ISO 8601 day of the week (Monday = 1, Sunday = 7). 7
w Number of the day of the week (Sunday = 0, Saturday = 6). 0
Hour
a "am" during the morning (00:00:00 → 11:59:59), "pm" otherwise (12:00:00 → 23:59:59). am
A Uppercase version of a above. AM
g Hour in 12-hour format, not zero-padded. 6
h Hour in 12-hour format, zero-padded. 06
G Hour in 24-hour format, not zero-padded. 6
H Hour in 24-hour format, zero-padded. 06
Minutes and seconds
i Minutes past the hour, zero-padded. 48
s Seconds past the minute, zero-padded. 40
U Seconds since January 1 1970 00:00:00 GMT. 1411282120
Timezone (as of 1.22wmf2)
e Timezone identifier. UTC
I Whether or not the date is in daylight savings time. 0
O Difference to Greenwich time (GMT) +0000
P Difference to Greenwich time (GMT), with colon +00:00
T Timezone abbreviation. UTC
Z Timezone offset in seconds. 0
Miscellaneous
t Number of days in the current month. 30
c ISO 8601 formatted date, equivalent to Y-m-dTH:i:s+00:00. 2014-09-21T06:48:40+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. Sun, 21 Sep 2014 06:48:40 +0000
Non-Gregorian calendars
Islamic
xmj Day of the month. 26
xmF Full month name. Dhu al-Qi'dah
xmn Month index. 11
xmY Full year. 1435
Iranian (Jalaly)
xij Day of the month. 30
xiF Full month name. Shahrivar
xin Month index. 6
xiY Full year. 1393
xiy 2-digit year. 93
Hebrew
xjj Day of the month. 26
xjF Full month name. Elul
xjt Number of days in month. 29
xjx Genitive form of the month name. Elul
xjn Month number. 12
xjY Full year. 5774
Thai solar
xkY Full year in Thai solar calendar.
Note Note: For years before 1941 the dates in Jan-Mar range are not calculated properly.
2557
Minguo/Juche year
xoY Full year. 103
Japanese nengo
xtY Full year. 平成26
Flags
xn Format the next numeric code as a raw ASCII number. In the Hindi language, {{#time:H, xnH}} produces 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}} → MMXIV
xh Format the next number as a Hebrew numeral. {{#time:xhY}} → ב'י"ד

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 }}2014-09-21
{{#time: [[Y]] m d }}2014 09 21
{{#time: [[Y (year)]] }}2014 (14UTCamSun, 21 Sep 2014 06:48:40 +0000)
{{#time: [[Y "(year)"]] }}2014 (year)
{{#time: i's" }}48'40"

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}}Sun, 21 Sep 2014 06:48:40 +0000
{{#time: r|+2 hours}}Sun, 21 Sep 2014 08:48:40 +0000
{{#time: r|now + 2 hours}}Sun, 21 Sep 2014 08:48:40 +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}}неділя
{{#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 }}1411282120
{{#time: r|@1411282120}}Sun, 21 Sep 2014 06:48:40 +0000

Warning Warning: 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 March 0100
(correct, no leap year), but
{{#time: r | 29 Feb 0100 }}Tue, 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

Year numbers 0-99 are interpreted as 2000-2069 and 1970-1999, even when written with leading zeros:
{{#time: d F Y | 1 Jan 0069 }}01 January 0069

{{#time: d F Y | 1 Jan 0070 }}01 January 0070

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 }}2014
Warning Warning: 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 }}2014 06 21 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 09 21 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 09 21 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 }}2014 09 21 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. Although 19:60:09 is not a valid time, 196009 is not interpreted as September 1960.

The function performs a certain amount of date mathematics:

{{#time: d F Y | January 0 2008 }}31 December 2007
{{#time: d F | January 32 }}Error: Invalid time.
{{#time: d F | February 29 2008 }}29 February
{{#time: d F | February 29 2007 }}01 March
{{#time:Y-F|now -1 months}}2014-August

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

Time Zone issue[edit | edit source]

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 }} ==> 2:48 AM

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 }} ==> 3:48 PM

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

  • {{#time:g:i A | -270 minutes }} ==> 2:18 AM
  • {{#time:g:i A | -16200 seconds }} ==> 2:18 AM

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


#timel[edit | edit source]

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 }}2014-09-21
{{#timel: Y-m-d }}2014-09-21
{{#time: Y F d h:i:s}}2014 September 21 06:48:40
{{#timel: Y F d h:i:s}}2014 September 21 06:48:40

#titleparts[edit | edit source]

This function separates a pagetitle 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 parameter is not specified, it defaults to "0", which returns all the segments from first segment (included). If the first segment 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
{{#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 number of segments effectively 'strips' segments from the end of the string. Negative values for first segment 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

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

Warning Warning: You can use #titleparts as a small "string parser & 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

Warning Warning: 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

Warning Warning:

{{#titleparts: [[page]]/123 | 1 | 2 }}page/123

Warning Warning: 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.

Strings parsing[edit | edit source]

An extension is available for additional string parsing, see Extension:StringFunctions.

General points[edit | edit source]

Substitution[edit | edit source]

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

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

Redirects[edit | edit source]

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

Escaping pipe characters in tables[edit | edit source]

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.

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

Description You type You get
Escaping pipe character as table row/column separator (most wikis)
{{!}}
|
Escaping pipe character as a plain character
&#124;
|


Stripping whitespace[edit | edit source]

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

For preventing trimming then- and else-parts, see m:Template:If. Some people use also <nowiki> </nowiki> instead of spaces.

See also[edit | edit source]