Help:Extension:ParserFunctions
The ParserFunctions extension provides eleven additional parser functions to supplement the "magic words", which are already present in MediaWiki. (It may be configured to provide additional parser functions for string handling; these string functions are documented elsewhere.) All the parser functions provided by this extension take the form:
{{#functionname: argument 1 | argument 2 | argument 3 ... }}
Contents
#expr[edit]
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. This function is also available in Scribunto via the mw.ext.ParserFunctions.expr
function.
{{#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{{#expr: 1 or -1 }}
→ 1{{#expr: -1 or 0 }}
→ 1{{#expr: 0 or 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 = }}
→ 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 9,539,829, where we want 9539829, 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}}
→ 9539929
The operator mod gives wrong results for some values of the second argument:
|
Rounding[edit]
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 because 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 | ||
{{#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 | ||
{{#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
|
Strings[edit]
Expressions only work with number-like values, they cannot compare strings or characters. #ifeq can be used instead.
{{#expr: "a" = "a" }}
→ Expression error: Unrecognized punctuation character """.{{#expr: a = a }}
→ Expression error: Unrecognized word "a".{{#ifeq: a | a | 1 | 0 }}
→ 1
#if[edit]
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:
→ no
| yes | 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
The last parameter (false) 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 a parameter of the enclosing #if function. Up to seven levels of nesting is possible, although that may depend on the wiki or a memory limit.
{{#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) }} }}
You can also use a parameter as the test string in your #if statement. You need to ensure you add the | (pipe symbol) after the name of the variable.
{{#if:{{{1|}}}|You entered text in variable 1|There is no text in variable 1}}
See Help:Parser functions in templates for more examples of this parser function.
#ifeq[edit]
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
Numerical comparisons with #ifeq and #switch are not equivalent to comparisons with #ifexpr :
because PHP compares two numbers of type integer, whereas:
because MediaWiki converts literal numbers in expressions to type float, which, for large integers like these, involves rounding. |
When used inside a parser function, any parser tags and other parser functions must be temporarily replaced with a unique code. This affects comparisons:
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. |
Literal comparisons to page-name magic words may fail depending on site configuration. For example, {{FULLPAGENAME}}, depending on wiki, may capitalize the first letter, and will replace all underscores with spaces, To work around this, apply the magic word to both parameters:
|
#iferror[edit]
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: {{#expr: . }} | error | correct }}
→ correct{{#iferror: <strong class="error">a</strong> | error | correct }}
→ error
#ifexpr[edit]
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]
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]
#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]
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]
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 with partial transclusion tags can effect a configuration file that enables an editor unfamiliar with template coding to view and edit configurable elements.
Default[edit]
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]
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
Comparison behavior[edit]
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
Numerical comparisons with #switch and #ifeq are not equivalent with comparisons in expressions (see also above):
|
Raw equal signs[edit]
"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=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.
Replacing #ifeq[edit]
#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]
Code | Description | Current output (Purge this page's cache to update) |
---|---|---|
Year | ||
Y |
4-digit year. | 2017 |
y |
2-digit year. | 17 |
L |
1 if it's a leap year, 0 if not. | 0 |
o [note 1] |
ISO-8601 year of the specified week.[note 2] | 2017[note 3] |
Month | ||
n |
Month index, not zero-padded. | 4 |
m |
Month index, zero-padded. | 04 |
M |
An abbreviation of the month name, in the site language. | Apr |
F |
The full month name in the site language. | April |
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)
(genitive)
|
Week | ||
W |
ISO 8601 week number, zero-padded. | 14 |
Day | ||
j |
Day of the month, not zero-padded. | 5 |
d |
Day of the month, zero-padded. | 05 |
z |
Day of the year (January 1 = 0). |
94 |
D |
An abbreviation for the day of the week. Rarely internationalized. | Wed |
l |
The full weekday name. Rarely internationalized. | Wednesday |
N |
ISO 8601 day of the week (Monday = 1, Sunday = 7). | 3 |
w |
Number of the day of the week (Sunday = 0, Saturday = 6). | 3 |
Hour | ||
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. | 5 |
h |
Hour in 12-hour format, zero-padded. | 05 |
G |
Hour in 24-hour format, not zero-padded. | 17 |
H |
Hour in 24-hour format, zero-padded. | 17 |
Minutes and seconds | ||
i |
Minutes past the hour, zero-padded. | 23 |
s |
Seconds past the minute, zero-padded. | 31 |
U |
Unix time. Seconds since January 1 1970 00:00:00 GMT. | 1491413011 |
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-d"T"H:i:s+00:00 . |
2017-04-05T17:23:31+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. |
Wed, 05 Apr 2017 17:23:31 +0000 |
Non-Gregorian calendars | ||
Islamic | ||
xmj |
Day of the month. | 8 |
xmF |
Full month name. | Rajab |
xmn |
Month index. | 7 |
xmY |
Full year. | 1438 |
Iranian (Jalaly) | ||
xit |
Number of days in the month. | 31 |
xiz |
Day of the year. | 15 |
xij |
Day of the month. | 16 |
xiF |
Full month name. | Farvardin |
xin |
Month index. | 1 |
xiY |
Full year. | 1396 |
xiy |
2-digit year. | 96 |
Hebrew | ||
xjj |
Day of the month. | 9 |
xjF |
Full month name. | Nisan |
xjt |
Number of days in month. | 30 |
xjx |
Genitive form of the month name. | Nisan |
xjn |
Month number. | 7 |
xjY |
Full year. | 5777 |
Thai solar | ||
xkY |
Full year in Thai solar calendar. |
2560 |
Minguo/Juche year | ||
xoY |
Full year. | 106 |
Japanese nengo | ||
xtY |
Full year. | 平成29 |
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}} → MMXVII |
xh |
Format the next number as a Hebrew numeral. | {{#time:xhY}} → ב'י"ז |
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:
- A backslash followed by a formatting character is interpreted as a single literal character
- 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 }}
→ 2017-04-05{{#time: [[Y]] m d }}
→ 2017 04 05{{#time: [[Y (year)]] }}
→ 2017 (17UTCpmWed, 05 Apr 2017 17:23:31 +0000){{#time: [[Y "(year)"]] }}
→ 2017 (year){{#time: i's" }}
→ 23'31"
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}}
→ Wed, 05 Apr 2017 17:23:31 +0000{{#time: r|+2 hours}}
→ Wed, 05 Apr 2017 19:23:31 +0000{{#time: r|now + 2 hours}}
→ Wed, 05 Apr 2017 19:23:31 +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}}
→ середа{{#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 }}
→ 1491413011{{#time: r|@1491413011}}
→ Wed, 05 Apr 2017 17:23:31 +0000
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.
Year numbers 0-99 are interpreted as 2000-2069 and 1970-1999, except when written in 4-digit format with leading zeros: |
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 }}
→ 2017
The fill-in feature is not consistent; some parts are filled in using the current values, others are not:
|
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 04 05 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 }}
→ 2017 04 05 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}}
→ 2017-March
The total length of the format strings of the calls of #time
is limited to 6000 characters[1].
Time Zone issue[edit]
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 }} ==> 1:23 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 }} ==> 2:23 AM
To workaround this issue, simply convert the time into minutes or seconds, like this:
- {{#time:g:i A | -270 minutes }} ==> 12:53 PM
- {{#time:g:i A | -16200 seconds }} ==> 12:53 PM
(Tim Starling, the developer of this function, provided the exact syntax for this solution.)
#timel[edit]
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 }}
→ 2017-04-05{{#timel: Y-m-d }}
→ 2017-04-05{{#time: Y F d H:i:s}}
→ 2017 April 05 17:23:31{{#timel: Y F d H:i:s}}
→ 2017 April 05 17:23:31
#titleparts[edit]
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
"
,"
, or"
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.
{{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'Or | 1 | 0 }}
are both returning1
; (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'Or = 1 | #default = 0 }}
are both returning1
; (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'Or | 1 | 0 }}
will all return1
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'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 stringCategory:Côte-d'Or
and not the expectedCategory:Côte-d'Or
, and that:{{PAGESINCAT: Côte-d'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'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:
- 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.
- 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
- 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 }}
→ '
- 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
You can use #titleparts as a small "string parser and converter", but consider that it returns the first substring capitalized:
If lower case is needed, use lc: function to control output:
You can prepend a 'dummy' slash at the beginning of the string to get the correct first substring capitalization (uppercase or lowercase). Use
|
Certain characters that are illegal in a page title will cause #titleparts to not parse the string:
|
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[edit]
Substitution[edit]
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.
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.
Redirects[edit]
Especially {{#time:
…|now-
…}} could be handy in redirects to pages including dates, but this does not work.
Escaping pipe characters in tables[edit]
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: |
.
Description | You type | You get |
---|---|---|
Escaping pipe character as table row/column separator |
{{!}} |
| |
Escaping pipe character as a plain character |
| |
| |
Stripping whitespace[edit]
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 }}foo
→ foobarfoofoo{{#if:||<nowiki /> bar <nowiki />}}foo
→ foo 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  
(breakable space) or
(non-breakable space), since they are not modified by the software:
<span style="white-space: pre;">foo{{#if:||   bar   }}foo</span>
→ foo bar foofoo{{#if:|| bar }}foo
→ foo bar foo
See also[edit]
- Help:Parser functions in templates
- Help:Calculation (on Meta)
- Help:Newlines and spaces on Meta
- Help:Comparison between ParserFunctions syntax and TeX syntax on Meta
- Help:Magic words
- Module:String obsoleting Extension:StringFunctions
- Extension:PhpTags