Help:Extension:ParserFunctions
The ParserFunctions extension provides 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 ... }}
Note: When you edit this page, you agree to release your contribution under the CC0. See Public Domain Help Pages for more info. |
#expr
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 sqrt
|
Binary | ^
|
* / div mod fmod
| |
+ -
| |
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 }}
Basic example
{{#expr: 1 + 1 }}
→ 2
The available operators are listed to the right, in order of precedence. See Manual:Expr parser function syntax 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 18,027,613, where we want 18027613, 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}}
→ 18027713
Rounding
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 (0.333333… → 0.33333) |
{{#expr: 1/6 round 5 }} |
0.16667 | Final digit is ≥ 5, so it is rounded up (0.166666… → 0.16667) |
{{#expr: 8.99999/9 round 5 }} |
1 | Again, the result is rounded up on the last digit, which results in additional rounding (0.999998… → 1.00000 → 1) |
{{#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 |
Rounding large numbers | ||
{{#expr: 1e-92 round 400 }} |
1.0E-92 | Rounding to a very large number leads to infinity. Hence, the original value without the infinity is given as the answer. |
{{#expr: 1e108 round 200 }} |
1.0E+108 | Same as above. |
Strings
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
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 (see #ifexpr for that):
{{#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.
(So that if the parameter does not have a value, it evaluates to an empty string instead of the string "{{{1}}}
".)
{{#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
This parser function compares two input strings, determines whether they are identical, and returns one of two strings based on the result.
If more comparisons and output strings are required, consider using #switch
.
{{#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
returning a valid number first)
As a practical example, consider an existing template Template:Timer
using the parser to choose between two standard times, short and long.
It takes the parameter as the first input to compare against the string "short" – there is no convention for the order, but it is simpler to read if the parameter goes first.
The template code is defined as:
{{#ifeq: {{{1|}}} | short | 20 | 40 }}
the following ensue:
{{timer|short}}
→ 20{{timer|20}}
→ 40{{timer}}
→ 40
#iferror
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
Some errors may cause a tracking category to be added, using {{#iferror:}}
will not suppress the addition of the category.
#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 of 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 }}
→
Boolean operators of equality or inequality operators are supported.
{{#ifexpr: 0 = 0 or 1 = 0 | yes}}
→ yes{{#ifexpr: 0 = 0 and 1 = 0 | | no}}
→ no{{#ifexpr: 2 > 0 or 1 < 0 | yes}}
→ yes{{#ifexpr: 2 > 0 and 1 > 0 | yes | no}}
→ yes
#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 | 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 }}
→ doesn't exist (because MediaWiki:Copyright has not 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
#ifexist:
is considered an "expensive parser function"; only a limited number of which can be included on any one page (including functions inside transcluded templates).
When this limit is exceeded, any further #ifexist:
functions automatically return false, whether the target page exists or not, and the page is categorized into Category:Pages with too many expensive parser function calls.
The name of the tracking category may vary depending on the content language of your wiki.
For some use cases it is possible to emulate the ifexist effect with css, by using the selectors a.new
(to select links to unexisting pages) or a:not(.new)
(to select links to existing pages).
Furthermore, since the number of expensive parser functions that can be used on a single page is controlled by $wgExpensiveParserFunctionLimit
, one can also increase the limit in LocalSettings.php if needed.
ifexist and wanted pages
A page that does not exist and is tested for using #ifexist will end up on the Wanted Pages. See task T14019 for the reason, and w:Template:Linkless exists for a workaround.
#rel2abs
This function converts a relative file path into an absolute filepath.
{{#rel2abs: path }}
{{#rel2abs: path | base path }}
Within the path
input, the following syntax is valid:
.
→ the current level..
→ go up one level/foo
→ go down one level into the subdirectory /foo
If the base path
is not specified, the full page name of the page will be used instead:
{{#rel2abs: /quok | Help:Foo/bar/baz }}
→ Help:Foo/bar/baz/quok{{#rel2abs: ./quok | Help:Foo/bar/baz }}
→ Help:Foo/bar/baz/quok{{#rel2abs: ../quok | Help:Foo/bar/baz }}
→ Help:Foo/bar/quok{{#rel2abs: ../. | Help:Foo/bar/baz }}
→ Help:Foo/bar
Invalid syntax, such as /.
or /./
, is ignored.
Since no more than two consecutive full stops are permitted, sequences such as these can be used to separate successive statements:
{{#rel2abs: ../quok/. | Help:Foo/bar/baz }}
→ Help:Foo/bar/quok{{#rel2abs: ../../quok | Help:Foo/bar/baz }}
→ Help:Foo/quok{{#rel2abs: ../../../quok | Help:Foo/bar/baz }}
→ quok{{#rel2abs: ../../../../quok | Help:Foo/bar/baz }}
→ Error: Invalid depth in path: "Help:Foo/bar/baz/../../../../quok" (tried to access a node above the root node).
For a similar group of functions see also Help:Magic words#URL data. Built-in parser functions include: 'localurl:', 'fullurl:', 'anchorencode:' etc.
#switch
See also: w:Help:Switch parser function
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 affect a configuration file that enables an editor unfamiliar with template coding to view and edit configurable elements.
Default
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 (an equals sign without {{}}
).
If it does, it will be treated as a case comparison, and no text will display if no cases match.
This is because the default value has not been defined (is empty).
If a case matches however, its associated string will be returned.
{{#switch: test | Bar | foo = Foo | baz = Baz }}
→{{#switch: test | foo = Foo | baz = Baz | B=ar }}
→{{#switch: test | test = Foo | baz = Baz | B=ar }}
→ Foo
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
.
The "#default =
" in the last parameter may be omitted in the above case.
Use with parameters
The function may be used with parameters as the test string.
In this case, it is not necessary to place the pipe after the parameter name, because it is very unlikely that you will choose to set a case to be the string "{{{parameter name}}}
".
(This is the value the parameter will default to if the pipe is absent and the parameter doesn't exist or have a value.
See Help:Parser functions in templates .)
{{#switch: {{{1}}} | foo = Foo | baz = Baz | Bar }}
In the above case, if {{{1}}}
equals foo
, the function will return Foo
.
If it equals baz
, the function will return Baz
.
If the parameter is empty or does not exist, the function will return Bar
.
As in the section above, cases can be combined to give a single result.
{{#switch: {{{1}}} | foo | zoo | roo = Foo | baz = Baz | Bar }}
Here, if {{{1}}}
equals foo
, zoo
or roo
, the function will return Foo
.
If it equals baz
, the function will return Baz
.
If the parameter is empty or does not exist, the function will return Bar
.
Additionally, the default result can be omitted if you do not wish to return anything if the test parameter value does not match any of the cases.
{{#switch: {{{1}}} | foo = Foo | bar = Bar }}
In this case, the function returns an empty string unless {{{1}}}
exists and equals foo
or bar
, in which case it returns Foo
or Bar
, respectively.
This has the same effect as declaring the default result as empty.
{{#switch: {{{1}}} | foo | zoo | roo = Foo | baz = Baz | }}
If for some reason you decide to set a case as "{{{parameter name}}}
", the function will return that case's result when the parameter doesn't exist or doesn't have a value.
The parameter would have to exist and have a value other than the string "{{{parameter name}}}
" to return the function's default result.
- (when
{{{1}}}
doesn't exist or is empty):{{#switch: {{{1}}} | {{{1}}} = Foo | baz = Baz | Bar }}
→ Foo
- (when
{{{1}}}
has the value "test
"):{{#switch: {{{1}}} | {{{1}}} = Foo | baz = Baz | Bar }}
→ Bar
- (when
{{{1}}}
has the value "{{{1}}}
"):{{#switch: {{{1}}} | {{{1}}} = Foo | baz = Baz | Bar }}
→ Foo
In this hypothetical case, you would need to add the pipe to the parameter ({{{1|}}}
).
Comparison behavior
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: 02 | +1 = one | +2 = two | three}}
→ two{{#switch: 100 | 1e1 = ten | 1e2 = hundred | other}}
→ hundred
{{#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
Raw equal signs
"Case" strings cannot contain raw equals signs. To work around this, use the {{=}} magic word, or replace equals sign with HTML code =
.
Example:
You type | You get |
---|---|
{{#switch: 1=2
| 1=2 = raw
| 1<nowiki>=</nowiki>2 = nowiki
| 1{{=}}2 = template
| default
}}
|
template |
{{#switch: 1=2
| 1=2 = html
| default
}}
|
html |
Replacing #ifeq
#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. deep nesting, linear:
{{#ifeq:{{{1}}}|condition1
|<!--then-->branch1
|<!--else-->{{#ifeq:{{{1}}}|condition2
|<!--then-->branch2
|<!--else-->{{#ifeq:{{{1}}}|condition3
|<!--then-->branch3
|<!--else-->branch4}}}}}}
On the other hand, the switch replacement could be complicated/impractical for IFs nested in both branches (shown with alternatives of indentation, indented on both sides), making full symmetrical tree:
{{#ifeq:{{{1}}}|condition1
|<!--then-->branch1t{{
#ifeq:{{{1}}}|condition2
|<!--then-->branch1t2t{{#ifeq:{{{1}}}|condition4|<!--then-->branch1t2t4t|<!--else-->branch1t2t4e}}
|<!--else-->branch1t2e{{#ifeq:{{{1}}}|condition5|<!--then-->branch1t2e5t|<!--else-->branch1t2e5e}}
}}
|<!--else-->branch1e{{#ifeq:{{{1}}}|condition3
|<!--then-->branch1e3t{{#ifeq:{{{1}}}|condition6|branch1e3t6t|branch1e3t6e}}
|<!--else-->branch1e3e{{
#ifeq:{{{1}}}|condition7
|branch1e3e7t
|branch1e3e7t
}}
}}
}}
#time
Code | Description | Current output (Purge this page's cache to update) |
---|---|---|
Year | ||
Y
|
4-digit year. | 2024 |
y
|
2-digit year. | 24 |
L
|
1 if it's a leap year, 0 if not. | 1 |
o [note 1]
|
ISO-8601 year of the specified week.[note 2] | 2024[note 3] |
Month | ||
n
|
Month index, not zero-padded. | 12 |
m
|
Month index, zero-padded. | 12 |
M
|
An abbreviation of the month name, in the site language. | Dec |
F
|
The full month name in the site language. | December |
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:{{#time:F Y|June 2010|pl}} → czerwiec 2010(nominative) {{#time:d xg Y|20 June 2010|pl}} → 20 czerwca 2010(genitive) |
Day of the month or the year | ||
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). Note: To get the ISO day of the year add 1. |
339 |
Week and day of the week | ||
W
|
ISO 8601 week number, zero-padded. | 49 |
N
|
ISO 8601 day of the week (Monday = 1, Sunday = 7). | 4 |
w
|
Number of the day of the week (Sunday = 0, Saturday = 6). | 4 |
D
|
An abbreviation for the day of the week. Rarely internationalized. | Thu |
l
|
The full weekday name. Rarely internationalized. | Thursday |
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. | 3 |
h
|
Hour in 12-hour format, zero-padded. | 03 |
G
|
Hour in 24-hour format, not zero-padded. | 3 |
H
|
Hour in 24-hour format, zero-padded. | 03 |
Minutes and seconds | ||
i
|
Minutes past the hour, zero-padded. | 04 |
s
|
Seconds past the minute, zero-padded. | 43 |
U
|
Unix time. Seconds since January 1 1970 00:00:00 GMT. | 1733367883 |
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. | 31 |
c
|
ISO 8601 formatted date, equivalent to Y-m-d"T"H:i:s+00:00 .
|
2024-12-05T03:04:43+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.
|
Thu, 05 Dec 2024 03:04:43 +0000 |
Non-Gregorian calendars | ||
Islamic | ||
xmj
|
Day of the month. | 3 |
xmF
|
Full month name. | Jumada al-thani |
xmn
|
Month index. | 6 |
xmY
|
Full year. | 1446 |
Iranian (Jalaly) | ||
xit
|
Number of days in the month. | 30 |
xiz
|
Day of the year. | 260 |
xij
|
Day of the month. | 15 |
xiF
|
Full month name. | Azar |
xin
|
Month index. | 9 |
xiY
|
Full year. | 1403 |
xiy
|
2-digit year. | 03 |
Hebrew | ||
xjj
|
Day of the month. | 4 |
xjF
|
Full month name. | Kislev |
xjt
|
Number of days in month. | 30 |
xjx
|
Genitive form of the month name. | Kislev |
xjn
|
Month number. | 3 |
xjY
|
Full year. | 5785 |
Thai solar | ||
xkY
|
Full year in Thai solar calendar. Note: For years before 1941 the dates in Jan-Mar range are not calculated properly. |
2567 |
Minguo/Juche year | ||
xoY
|
Full year. | 113 |
Japanese nengo | ||
xtY
|
Full year. | 令和6 |
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 (up to 3,000 in pre MediaWiki 1.20). |
{{#time:xrY}} → MMXXIV
|
xh
|
Format the next number as a Hebrew numeral. | {{#time:xhY}} → ב'כ"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 }}
{{#time: format string | date/time object | language code | local }}
The list of accepted formatting codes is given in the table to the right. Any character in the formatting string that is not recognized is passed through unaltered; this applies also to blank spaces (the system does not need them for interpreting the codes). If no character is recognized in the formatting string, and the date/time object is without error, then the formatting string is returned as output. 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".
As the list of formatting codes continues to evolve (with the support of new calendars, or of new date fields computed and formatted differently), you should escape all literal characters (not just ASCII letters currently used by formatting codes) that need to be passed through unaltered.
Unfortunately, for now, the ASCII single quote is still not recognized as a simple alternative for marking literal text to the currently supported ASCII double quotes (for example, double quotes are mandatory for in other uses like the delimitation of string values in JSON, C, C++...) and backslashes (which have to be escaped as well in string constants used by many languages, including JSON, C, C++, PHP, JavaScript, Lua). So you still cannot embed any literal double quote without escaping it with a backslash (or you can use other curly, angular or square quotation marks instead).
{{#time: Y-m-d }}
→ 2024-12-05{{#time: [[Y]] m d }}
→ 2024 12 05{{#time: [[Y (year)]] }}
→ 2024 (24UTCamThu, 05 Dec 2024 03:04:43 +0000){{#time: [[Y "(year)"]] }}
→ 2024 (year){{#time: i's" }}
→ 04'43"
The date/time object
can be in any format accepted by PHP's strtotime() function.
Absolute (e.g. 20 December 2000
), relative (e.g. +20 hours
), and combined times (e.g. 30 July +1 year
) are accepted.
{{#time: r|now}}
→ Thu, 05 Dec 2024 03:04:44 +0000{{#time: r|+2 hours}}
→ Thu, 05 Dec 2024 05:04:44 +0000{{#time: r|now + 2 hours}}
→ Thu, 05 Dec 2024 05:04:44 +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}}
→ четвер{{#time:d xg Y|20 June 2010|pl}}
→ 20 czerwca 2010
The local
parameter specifies if the date/time object refers to the local timezone or to UTC.
This is a boolean parameters: its value is determined by casting the value of the argument (see the official PHP documentation for details on how string are cast to boolean values).
$wgLocaltimezone
is set to UTC
, there is no difference in the output when local
is set to true
or false
.See the following examples for details:
{{#time: Y F d H:i:s|now|it|0}}
→ 2024 dicembre 05 03:04:44{{#time: Y F d H:i:s|now|it|1}}
→ 2024 dicembre 05 03:04:44{{#time: Y F d H:i:s|+2 hours||0}}
→ 2024 December 05 05:04:44{{#time: Y F d H:i:s|+2 hours||1}}
→ 2024 December 05 05:04:44
{{#time:c|2019-05-16T17:05:43+02:00|it}}
→ 2019-05-16T15:05:43+00:00{{#time:c|2019-05-16T17:05:43+02:00|it|0}}
→ 2019-05-16T15:05:43+00:00{{#time:c|2019-05-16T17:05:43+02:00|it|true}}
→ 2019-05-16T15:05:43+00:00
If you've calculated a Unix timestamp, you may use it in date calculations by pre-pending an @
symbol.
{{#time: U | now }}
→ 1733367884{{#time: r | @1733367883 }}
→ Thu, 05 Dec 2024 03:04:43 +0000
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 }}
→ 2024
A four-digit number is always interpreted as a year, never as hours and minutes:[1]
{{#time: Y m d H:i:s | 1959 }}
→ 1959 12 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 }}
→ 2024 12 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}}
→ 2024-November
The total length of the format strings of the calls of #time
is limited to 6000 characters[2].
Time Zone issue
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 }}
→ 11:04 PM
However, India is on a +5.5 hours time offset from UTC, and thus using its time zone will not normally allow the correct calculation of a relative time zone offset. Here's what happens:
{{#time:g:i A | +5.5 hours }}
→ 3:04 AM
To workaround this issue, simply convert the time into minutes or seconds, like this:
{{#time:g:i A | +330 minutes }}
→ 8:34 AM{{#time:g:i A | +19800 seconds }}
→ 8:34 AM
(Tim Starling, the developer of this function, provided the exact syntax for this solution.)
#time format like in signatures
Sometimes it is useful to construct a timestamp, which looks like the automatic timestamp generated by signatures in discussions on talk pages. On an English-language wiki, it can be created with:
{{#timel:H:i, j xg Y (e)|+330 minutes}}
→ 08:34, 5 December 2024 (UTC)
#timel
This function is a syntactic shortcut that operates identically to {{#time: ... }}
with the local
parameter set to true
, so it always uses the preferred time zone of the user or the configured time zone of the wiki (as set in $wgLocaltimezone )
Syntax of the function is:
{{#timel: format string }}
{{#timel: format string | date/time object }}
{{#timel: format string | date/time object | language code }}
$wgLocaltimezone
is set to UTC
, there is no difference in the output when local
is set to true
or false
For instance, see the following examples:
{{#time:c|now|it}}
→ 2024-12-05T03:04:44+00:00{{#time:c|now|it|0}}
→ 2024-12-05T03:04:44+00:00{{#time:c|now|it|1}}
→ 2024-12-05T03:04:44+00:00{{#timel:c|now|it}}
→ 2024-12-05T03:04:44+00:00
#timef
This function formats a date using a standard format for the selected language, as defined in $dateFormats
(see task T223772).
{{#timef: date/time object }}
{{#timef: date/time object | format type }}
{{#timef: date/time object | format type | language code }}
The format of the date/time object is the same as for #time. If it is empty, the time when the page was rendered is used.
The format type may be one of:
time
- Only the time is shown.
date
- Only the date is shown.
both
- Both the time and date are shown.
pretty
- Only the date is shown, using an abbreviated format which does not include the year. Not all languages support this; if it is not supported, the "date" format is used.
If the format type is not specified, both the time and date will be show, as if both
were specified.
If the language code is not specified, the page's content language is used.
Using #timef
instead of #time
allows templates to more easily support multiple languages, since different languages have different ways to format dates.
In English, the order of the day and month is controlled by $wgAmericanDates
.
Examples:
{{#timef:now|both|en}}
→ 03:04, 5 December 2024{{#timef:now|both|ja}}
→ 2024年12月5日 (木) 03:04{{#timef:now|pretty|en}}
→ 5 December{{#timef:now|pretty|pl}}
→ 5 grudnia{{#timef:|time}}
→ 03:04
#timefl
This function is the same as #timef except that it uses the local timezone of the wiki as configured in $wgLocaltimezone .
{{#timefl: date/time object }}
{{#timefl: date/time object | format type }}
{{#timefl: date/time object | format type | language code }}
#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 | segment to start at }}
If the number of segments to return parameter is not specified, it defaults to "0", which returns all the segments from the segment to start at to the end (included). If the segment to start at 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 | 3 }}
→ baz/quok{{#titleparts: Talk:Foo/bar/baz/quok | 3 | 2 }}
→ bar/baz/quok{{#titleparts: Talk:Foo/bar/baz/quok | | 2 }}
→ bar/baz/quok{{#titleparts: Talk:Foo/bar/baz/quok | | 5 }}
→
Negative values are accepted for both values. Negative values for the number of segments to return parameter effectively 'strips' segments from the end of the string. Negative values for the first segment to return translates to "start with this segment counting from the right":
{{#titleparts: Talk:Foo/bar/baz/quok | -1 }}
→ Talk:Foo/bar/baz Strips one segment from the end of the string. See also {{BASEPAGENAME}}.{{#titleparts: Talk:Foo/bar/baz/quok | -4 }}
→ Strips all 4 segments from the end of the string{{#titleparts: Talk:Foo/bar/baz/quok | -5 }}
→ Strips 5 segments from the end of the string (more than exist){{#titleparts: Talk:Foo/bar/baz/quok | | -1 }}
→ quok Returns last segment. See also {{SUBPAGENAME}}.{{#titleparts: Talk:Foo/bar/baz/quok | -1 | 2 }}
→ bar/baz Strips one segment from the end of the string, then returns the second segment and beyond{{#titleparts: Talk:Foo/bar/baz/quok | -1 | -2 }}
→ baz Start copying at the second last element; strip one segment from the end of the string
Before processing, the pagename parameter is HTML-decoded: if it contains some standard HTML character entities, they will be converted to plain characters (internally encoded with UTF-8, i.e. the same encoding as in the MediaWiki source page using this parser function).
- For example, any occurrence 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
- 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
{{#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
String functions
The ParserFunctions extension optionally defines various string functions if $wgPFEnableStringFunctions
is set to true
:
#len
#pos
#rpos
#sub
#count
#replace
#explode
#urldecode
See the dedicated subpage for documentation, and Manual:Performing string operations with parser functions for examples.
Here is a short overview of Module:String functions:
- #len (length of string):
{{#invoke:String|len|target_string}}
- #sub (substring):
{{#invoke:String|sub|target_string|start_index|end_index}}
- #match:
{{#invoke:String|match|source_string|pattern_string|start_index|match_number|plain_flag|nomatch_output}}
- #pos (position of target):
{{#invoke:String|pos|target_string|index_value}}
- #find:
{{#invoke:String|find|source_string|target_string|start_index|plain_flag}}
- #replace:
{{#invoke:String|replace|source_str|pattern_string|replace_string|replacement_count|plain_flag}}
- #rep (repeat):
{{#invoke:String|rep|source|count}}
- #escapePattern:
{{#invoke:String|escapePattern|pattern_string}}
- #count:
{{#invoke:String|count|source_str|pattern_string|plain_flag}}
- #join:
{{#invoke:String|join|separator|string1|string2|...}}
General points
Substitution
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.
Substitution does not work within <ref>
…</ref>
; you can use {{subst:#tag:ref|
…}}
for this purpose.
Redirects
Especially {{#time:
…|now-
…}} could be handy in redirects to pages including dates, but this does not work.
Escaping pipe characters in tables
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
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/> bar <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
Beware that not all parameters are created equal. In ParserFunctions, whitespace at the beginning and end is always stripped. In templates , whitespace at the beginning and end is stripped for named parameters and named unnamed parameters but not from unnamed parameters:
Other parser functions
Case conversion functions
- Lowercase:
"{{lc: AbC}}"
→ "abc" [1] - Uppercase:
"{{uc: AbC}}"
→ "ABC" [2] - Lowercase first character:
"{{lcfirst: AbC}}"
→ "abC" [3] - Uppercase first character:
"{{ucfirst: abc}}"
→ "Abc" [4]
Encoding functions
- URL encoding:
"{{urlencode: AbC
dEf ghi}}"
renders as
"AbC%0AdEf+ghi"
So inner new lines convert into %0A, and inner spaces convert into +.
Anchor encoding
{{anchorencode: AbC dEf ghi}}
renders as
AbC_dEf_ghi
Padding functions
Formatting functions
"{{formatnum: 1234567890}}"
gives "1,234,567,890" [7]
See also
- Help:Parser functions in templates
- Parser extension tags
- Parser function hooks - an (incomplete) list of parser functions added by core and extensions.
- Manual:Newlines and spaces
- Manual:Converting between ParserFunctions syntax and TeX syntax
- Help:Magic words
- Module:String obsoleting Extension:StringFunctions
- Extension:PhpTags
- Parser functions for Wikibase (the extensions that enables Wikidata): How to use data on Wikimedia projects#Parser function
References
- ↑ Prior to r86805 in 2011 this was not the case.
- ↑ ParserFunctions.php at phabricator.wikimedia.org