Help:Extension:ParserFunctions - MediaWiki

Help:Extension:ParserFunctions - MediaWiki
Jump to content
From mediawiki.org
Translate this page
Languages:
Bahasa Indonesia
Lëtzebuergesch
Nederlands
Tiếng Việt
Türkçe
Yorùbá
català
galego
italiano
magyar
polski
português
português do Brasil
slovenčina
slovenščina
suomi
svenska
čeština
беларуская (тарашкевіца)
български
русский
українська
հայերեն
العربية
فارسی
پښتو
मराठी
हिन्दी
বাংলা
ગુજરાતી
தமிழ்
ไทย
中文
한국어
Note:
When you edit this page, you agree to release your contribution under the
CC0
. See
Public Domain Help Pages
for more info.
Magic words
Magic words
Additional parser functions
ParserFunctions
String functions module
Scribunto
Navbox of parser functions
The
ParserFunctions
extension provides additional
parser functions
to supplement those already present in MediaWiki core.
(See
Help:Magic words
.)
(It
may
be configured to provide additional parser functions for string handling; these are only available to non-Wikimedia Foundation wikis hence these string functions are documented
elsewhere
.)
All the parser functions provided by this extension take the form:
{{
#functionname
argument 1
argument 2
argument 3
... }}
#expr
For a more
in-depth manual
on the finer points of how the expression evaluator works, including some additional operators not covered here, see:
Manual:Expr parser function syntax
Type
Operators
Grouping (parentheses)
( )
Numbers
1234.5
(2.718)
pi
(3.142)
binary operator
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
}}
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
}}
{{
#expr
1 and 0
}}
{{
#expr
1 or -1
}}
{{
#expr
-1 or 0
}}
{{
#expr
0 or 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
}}
{{
#expr
-1
}}
-1
{{
#expr
+ 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 18,352,186, where we want 18352186, 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
}}|
}}
+100
}}
18352286
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 (ex. test whether current date and time is after some other date and time), first convert the time to number of seconds after January 1, 1970 (UTC) using {{#time: xNU }}, then you can simply add and subtract dates as numbers.
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 }}
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''' }}
Down to the
nearest
integer, which is zero
{{#expr: 1/2 '''round 0''' }}
Up to the nearest integer, which is one
{{#expr: 3/4 '''round 0''' }}
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''')''' }}
Up to the next
larger
integer, which is one
{{#expr: '''floor('''1/3''')''' }}
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
Warning:
Interpreted as
(ceil
/3, not
ceil(
1/3
, as you might expect
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
}}
#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
}}
yes
{{
#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
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:
{{{
}}}
|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
equal
not equal
}}
equal
{{
#ifeq
-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
{{{
|}}}
short
20
40
}}
the following ensue:
{{
timer
short
}}
20
{{
timer
20
}}
40
{{
timer
}}
40
Warning:
When used inside a parser function, any parser tags and other parser functions must be temporarily replaced with
a unique code
. This affects comparisons:
{{#ifeq: foo | foo | equal | not equal}}
not equal
{{#ifeq: $foo$ | $foo$ | 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:
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:
{{
#ifeq
{{
FULLPAGENAME
L'Aquila
}}
{{
FULLPAGENAME
}}
equal
not equal
}}
equal
#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
}}
{{
#iferror
{{
#expr
1 + X
}}
error
}}
error
{{
#iferror
{{
#expr
1 + 2
}}
}}
{{
#iferror
{{
#expr
1 + X
}}
}}
{{
#iferror
{{
#expr
}}
error
correct
}}
correct
{{
#iferror
strong
class
"error"
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
}}
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
Warning:
The results of numerical comparisons with
#ifexpr
do not always match those of
#ifeq
and
#switch
. These latter two are more accurate than
#ifexpr
, and so may not return equivalent results.
Consider these comparisons with the final digit changed:
{{
#ifeq
12345678901234567
12345678901234568
equal
not equal
}}
not equal
{{
#switch
12345678901234567
12345678901234568
equal
not equal
}}
not equal
Because PHP used in
#ifeq
and
#switch
compares two numbers of type integer, it returns the expected result correctly.
Whereas with
#ifexpr
and the same numbers:
{{
#ifexpr
12345678901234567 = 12345678901234568
equal
not equal
}}
equal
With the different digit, the result of equal is actually incorrect.
This behaviour in
#ifexpr
is caused because MediaWiki converts literal numbers in expressions to type float, which, for large integers like these, involves rounding.
#ifexist
See
Manual:Checking for page existence
for other methods of checking if a page exists with different limitations
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)
Prior to MediaWiki 1.45 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 nonexistent 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
Prior to MediaWiki 1.45, a page that does not exist and is tested for using #ifexist will end up on the
Wanted Pages
See
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 be used to make a configuration file that enables editors 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
ar
}}
{{
#switch
test
test
Foo
baz
Baz
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
{{{
}}}
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
{{{
}}}
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
{{{
}}}
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
{{{
}}}
foo
zoo
roo
Foo
baz
Baz
}}
If 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
{{{
}}}
{{{
}}}
= Foo
baz
Baz
Bar
}}
Foo
(when
{{{1}}}
has the value "
test
"):
{{
#switch
{{{
}}}
{{{
}}}
= Foo
baz
Baz
Bar
}}
Bar
(when
{{{1}}}
has the value "
{{{1}}}
"):
{{
#switch
{{{
}}}
{{{
}}}
= 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
one
two
three
}}
three
{{
#switch
{{
#expr
0 + 1
}}
one
two
three
}}
one
{{
#switch
02
+1
one
+2
two
three
}}
two
{{
#switch
100
1e1
ten
1e2
hundred
other
}}
hundred
{{
#switch
}}
{{
#switch
}}
case
string may be empty:
{{
#switch
Nothing
foo
Foo
Something
}}
Nothing
Once a match is found, subsequent
cases
are ignored:
{{
#switch
Foo
Bar
Baz
}}
Bar
Warning:
Numerical comparisons with
#switch
and
#ifeq
are not equivalent to comparisons in expressions (see also above):
{{
#switch
12345678901234567
12345678901234568
}}
{{
#ifexpr
12345678901234567 = 12345678901234568
}}
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
2 = raw
nowiki
nowiki
2 = nowiki
{{
}}
2 = template
default
}}
template
{{
#switch
1=2
1=2
html
default
}}
html
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
w:Template:BOTREQ
Replacing #ifeq
#switch
can be used to reduce
expansion depth
For example:
{{
#switch
:{{{
}}}
condition1
branch1
condition2
branch2
condition3
branch3
branch4
}}
is equivalent to
{{
#ifeq
:{{{
}}}|
condition1
branch1
|{{
#ifeq
:{{{
}}}|
condition2
branch2
|{{
#ifeq
:{{{
}}}|
condition3
branch3
branch4
}}}}}}
i.e. deep nesting, linear:
{{
#ifeq
:{{{
}}}|
condition1

branch1

{{
#ifeq
:{{{
}}}|
condition2

branch2

{{
#ifeq
:{{{
}}}|
condition3

branch3

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
:{{{
}}}|
condition1

branch1t
{{
#ifeq
:{{{
}}}|
condition2

branch1t2t
{{
#ifeq
:{{{
}}}|
condition4

branch1t2t4t

branch1t2t4e
}}

branch1t2e
{{
#ifeq
:{{{
}}}|
condition5

branch1t2e5t

branch1t2e5e
}}
}}

branch1e
{{
#ifeq
:{{{
}}}|
condition3

branch1e3t
{{
#ifeq
:{{{
}}}|
condition6
branch1e3t6t
branch1e3t6e
}}

branch1e3e
{{
#ifeq
:{{{
}}}|
condition7
branch1e3e7t
branch1e3e7t
}}
}}
}}
#time
Code
Description
Current output
Purge this page's cache
to update)
Year
4-digit year.
2026
2-digit year.
26
1 if it's a leap year, 0 if not.
note 1
ISO-8601 year of the specified week.
note 2
2026
note 3
Requires PHP 5.1.0 and newer and
rev:45208
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.
Will output literal
if note 1 not fulfilled.
Month
Month index, not zero-padded.
Month index, zero-padded.
04
An abbreviation of the month name, in the site language.
Apr
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:
{{#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
Day of the month, not zero-padded.
23
Day of the month, zero-padded.
23
Day of the year (January 1 = 0).
Note:
To get the ISO day of the year add 1.
112
Week and day of the week
ISO 8601 week number, zero-padded.
17
ISO 8601 day of the week (Monday = 1, Sunday = 7).
Number of the day of the week (Sunday = 0, Saturday = 6).
An abbreviation for the day of the week. Rarely internationalized.
Thu
The full weekday name. Rarely internationalized.
Thursday
Hour
"am" during the morning (00:00:00 → 11:59:59), "pm" otherwise (12:00:00 → 23:59:59).
pm
Uppercase version of
above.
PM
Hour in 12-hour format, not zero-padded.
Hour in 12-hour format, zero-padded.
05
Hour in 24-hour format, not zero-padded.
17
Hour in 24-hour format, zero-padded.
17
Minutes and seconds
Minutes past the hour, zero-padded.
37
Seconds past the minute, zero-padded.
26
Unix time
. Seconds since January 1 1970 00:00:00 GMT.
1776965846
Timezone (as of
1.22wmf2
Timezone identifier.
UTC
Whether or not the date is in daylight savings time.
Difference to Greenwich time (GMT)
+0000
Difference to Greenwich time (GMT), with colon
+00:00
Timezone abbreviation.
UTC
Timezone offset in seconds.
Miscellaneous
Number of days in the current month.
30
ISO 8601 formatted date, equivalent to
Y-m-d"T"H:i:s+00:00
2026-04-23T17:37:26+00:00
RFC 5322
formatted date, equivalent to
D, j M Y H:i:s +0000
, with weekday name and month name not internationalized.
Thu, 23 Apr 2026 17:37:26 +0000
Non-Gregorian calendars
Islamic
xmj
Day of the month.
xmF
Full month name.
Dhu al-Qi'dah
xmn
Month index.
11
xmY
Full year.
1447
Iranian (Jalaly)
xit
Number of days in the month.
31
xiz
Day of the year.
33
xij
Day of the month.
xiF
Full month name.
Ordibehesht
xin
Month index.
xiY
Full year.
1405
xiy
2-digit year.
05
Hebrew
xjj
Day of the month.
xjF
Full month name.
Iyar
xjt
Number of days in month.
29
xjx
Genitive form of the month name.
Iyar
xjn
Month number.
xjY
Full year.
5786
Thai solar
xkY
Full year in
Thai solar calendar
Note:
For years before 1941 the dates in Jan-Mar range are not
calculated
properly.
2569
Minguo/Juche year
xoY
Full year.
115
Japanese nengo
xtY
Full year.
令和8
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
}}
→ MMXXVI
xh
Format the next number as a Hebrew numeral.
{{
#time
xhY
}}
→ ב'כ"ו
This parser function takes a date and/or time (in the Gregorian calendar) and formats it according to the syntax given. A date/time object can be specified; the default is the value of the
magic word
{{
CURRENTTIMESTAMP
}}
– that is, the time the page was last rendered into HTML.
{{#time:
format string
}}
{{#time:
format string
date/time object
}}
{{#time:
format string
date/time object
language code
}}
{{#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
}}
2026-04-23
{{
#time
[[
]]
m d
}}
2026
04 23
{{
#time
[[
Y (year)
]]
}}
2026 (26UTCpmThu, 23 Apr 2026 17:37:26 +0000)
{{
#time
[[
Y "(year)"
]]
}}
2026 (year)
{{
#time
i's"
}}
37'26"
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
now
}}
Thu, 23 Apr 2026 17:37:27 +0000
{{
#time
+2 hours
}}
Thu, 23 Apr 2026 19:37:27 +0000
{{
#time
now + 2 hours
}}
Thu, 23 Apr 2026 19:37:27 +0000
{{
#time
20 December 2000
}}
Wed, 20 Dec 2000 00:00:00 +0000
{{
#time
December 20, 2000
}}
Wed, 20 Dec 2000 00:00:00 +0000
{{
#time
2000-12-20
}}
Wed, 20 Dec 2000 00:00:00 +0000
{{
#time
2000 December 20
}}
Error: Invalid time.
{{
#time
last tuesday
}}
Tue, 21 Apr 2026 00:00:00 +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
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).
Please note that, if the variable
$wgLocaltimezone
is set to
UTC
, there is no difference in the output when
local
is set to
true
or
false
See the following examples for details:
{{
#time
Y F d H:i:s
now
it
}}
2026 aprile 23 17:37:27
{{
#time
Y F d H:i:s
now
it
}}
2026 aprile 23 17:37:27
{{
#time
Y F d H:i:s
+2 hours
||
}}
2026 April 23 19:37:27
{{
#time
Y F d H:i:s
+2 hours
||
}}
2026 April 23 19:37:27
{{
#time
2019-05-16T17:05:43+02:00
it
}}
2019-05-16T15:05:43+00:00
{{
#time
2019-05-16T17:05:43+02:00
it
}}
2019-05-16T15:05:43+00:00
{{
#time
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
now
}}
1776965847
{{#time: r | @1776965846 }}
Thu, 23 Apr 2026 17:37:26 +0000
Warning:
Without the
prefix before numeric timestamp values, the result is an error most of the time, or is an unexpected value:
{{
#time
1970-01-01 00:16:39
}}
Thu, 01 Jan 1970 00:16:39 +0000
{{
#time
1970-01-01 00:16:39
}}
999
{{#time: r | @999 }}
Thu, 01 Jan 1970 00:16:39 +0000
(correct)
{{#time: r | 999 }}
Error: Invalid time.
(unsupported year format)
{{
#time
1970-01-01 00:16:40
}}
Thu, 01 Jan 1970 00:16:40 +0000
{{
#time
1970-01-01 00:16:40
}}
1000
{{#time: r | @1000 }}
Thu, 01 Jan 1970 00:16:40 +0000
(correct)
{{#time: r | 1000 }}
Wed, 23 Apr 1000 00:00:00 +0000
(interpreted as a year with current month and day of the month)
{{
#time
1970-01-01 02:46:39
}}
Thu, 01 Jan 1970 02:46:39 +0000
{{
#time
1970-01-01 02:46:39
}}
9999
{{#time: r | @9999 }}
Thu, 01 Jan 1970 02:46:39 +0000
(correct)
{{#time: r | 9999 }}
Fri, 23 Apr 9999 00:00:00 +0000
(interpreted as a year with current month and day of the month)
{{
#time
1970-01-01 02:46:40
}}
Thu, 01 Jan 1970 02:46:40 +0000
{{
#time
1970-01-01 02:46:40
}}
10000
{{#time: r | @10000 }}
Thu, 01 Jan 1970 02:46:40 +0000
(correct)
{{#time: r | 10000 }}
Error: Invalid time.
(unsupported year format)
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
29 Feb 0100
}}
Mon, 01 Mar 0100 00:00:00 +0000
(wrong, even if 100 is interpreted as 2000, because that is a leap year)
{{
#time
d F Y
15 April 10000
}}
Error: Invalid time.
{{
#time
10000-4-15
}}
Sat, 15 Apr 2000 10:00:00 +0000
Year numbers 0-99 are interpreted as 2000-2069 and 1970-1999, except when written in 4-digit format with leading zeros:
{{
#time
d F Y
1 Jan 6
}}
01 January 2006
{{
#time
d F Y
1 Jan 06
}}
01 January 2006
{{
#time
d F Y
1 Jan 006
}}
01 January 2006
{{
#time
d F Y
1 Jan 0006
}}
01 January 0006
(4-digit format)
The weekday is supplied for the years 100-110 and from 1753, for the years 111-1752 the r-output shows "Unknown" and the l-output "<>". As a consequence, the r-output is not accepted as input for these years.
Full or partial absolute dates can be specified; the function will "fill in" parts of the date that are not specified using the
current
values:
{{
#time
January 1
}}
2026
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
}}
2026 06 23 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 04 23 00:00:00
Gives the start of the day, but the current day of the year.
There's exception case of the filled day:
{{
#time
Y m d H:i:s
June 2003
}}
2003 06 01 00:00:00
Gives the start of the day and the start of the month.
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 23 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
}}
2026 04 23 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
}}
2026-March
The total length of the format strings of the calls of
#time
is limited to 6000 characters.
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
}}
→ 1:37 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
}}
→ 5:37 PM
To workaround this issue, simply convert the time into minutes or seconds, like this:
{{
#time
g:i A
+330 minutes
}}
→ 11:07 PM
{{
#time
g:i A
+19800 seconds
}}
→ 11:07 PM
(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
}}
→ 23:07, 23 April 2026 (UTC)
#timel
This function is identical to
{{#time: ... }}
with the
local
parameter set to
true
, so it always uses the local time of the wiki (as set in
$wgLocaltimezone
).
Syntax of the function is:
{{#timel:
format string
}}
{{#timel:
format string
date/time object
}}
{{#timel:
format string
date/time object
language code
}}
Please note that, if the variable
$wgLocaltimezone
is set to
UTC
, there is no difference in the output when
local
is set to
true
or
false
Example of the use of #time and #timel parser functions from a server where the timezone is not UTC
For instance, see the following examples:
{{
#time
now
it
}}
2026-04-23T17:37:27+00:00
{{
#time
now
it
}}
2026-04-23T17:37:27+00:00
{{
#time
now
it
}}
2026-04-23T17:37:27+00:00
{{
#timel
now
it
}}
2026-04-23T17:37:27+00:00
Warning Example from
Warning:
Be aware that U for both time and timel will return the same number of seconds since 1970-01-01 00:00:00 UTC on Wikipedias with different timezones than UTC (formerly known as GMT)
Unix time. Seconds since January 1 1970 00:00:00 GMT.
Timezone offset in seconds.
{{
#time
}}
1776965846
{{
#timel
}}
1776965846
{{
#time
}}
{{
#timel
}}
#timef
This function formats a date using a standard format for the selected language, as defined in
$dateFormats
(see
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
}}
17:37, 23 April 2026
{{
#timef
now
both
ja
}}
2026年4月23日 (木) 17:37
{{
#timef
now
pretty
en
}}
23 April
{{
#timef
now
pretty
pl
}}
23 kwietnia
{{
#timef
:|
time
}}
17:37
#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.
Some magic keywords or parser functions of MediaWiki (such as
{{
PAGENAME
}}
and similar) are known to return strings that are needlessly HTML-encoded, even if their own input parameter was not HTML-encoded:
The titleparts parser function can then be used as a workaround, to convert these returned strings so that they can be processed correctly by some other parser functions also taking a page name in parameter (such as
{{
PAGESINCAT:
}}
) but which are still not working properly with HTML-encoded input strings.
This may not be a suitable workaround for you if a pagename
is allowed to start with a lowercase character
: the titleparts parser function will always capitalize the first letter.
For example, if the current page is
Category:Côte-d'Or
, then:
{{
#ifeq
{{
FULLPAGENAME
}}
Category:Côte-d'Or
}}
, and
{{
#ifeq
{{
FULLPAGENAME
}}
Category:Côte-d
'
Or
}}
are both returning
; (the #ifeq parser function does perform the HTML-decoding of its input parameters).
{{
#switch
{{
FULLPAGENAME
}}
Category:Côte-d'Or
#default
}}
, and
{{
#switch
{{
FULLPAGENAME
}}
Category:Côte-d'Or
#default
}}
are both returning
; (the #switch parser function does perform the HTML-decoding of its input parameters).
{{
#ifexist
{{
FULLPAGENAME
}}
}}
{{
#ifexist
Category:Côte-d'Or
}}
, or even
{{
#ifexist
Category:Côte-d
'
Or
}}
will all return
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 string
Category:Côte-d'Or
and not the expected
Category: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/
/aa/bb/cc/dd/ee| 1 | 25 }} | 1 | 2}}
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:
{{#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
instead of
for
first segment to return
{{#titleparts:
one
/two/three/four|1|
}}
one
{{#titleparts:
One
/two/three/four|1|
}}
One
Warning:
Certain characters that are
illegal in a page title
will cause #titleparts to not parse the string:
{{
#titleparts
{one/two}
}}
{one/two}
. Does not produce the expected:
{one
{{
#titleparts
[[
page
]]
/123
}}
page
/123
. Does not work because brackets are illegal in page titles and this parser function does not process links embedded in its input
pagename
parameter, even when they use the MediaWiki syntax, or any other HTML or MediaWiki tags.
{{
#titleparts
red/#00FF00/blue
}}
→ "". Does not work because "#" is also illegal in page titles.
Warning:
If any part of the title is just "
" or "
..
", #titleparts will not parse the string:
{{
#titleparts
one/./three
}}
one/./three
. The whole string is returned. It does not produce the expected:
one
Warning:
This function does not degrade gracefully if the input exceeds 255 bytes in UTF-8. If the input string is 256 bytes or more, the whole string is returned.
String functions
Main page:
Extension:ParserFunctions/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.
Warning:
In 2013, it was decided that
these functions will
never
be enabled on any Wikimedia wiki
, because they are inefficient when used on a large scale (see
phab:T8455
for some history).
These functions do
NOT
work on Wikimedia wikis!
If you are here to write something on a Wikimedia project, you are looking for something else: if your home wiki has string functions, it probably uses
Lua
. For example, the English Wikipedia uses
Module:String
, which does some of the same things with wildly different syntax. There are also individual
String-handling templates
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.
Warning:
The results of substituted parser functions are undefined if the expressions contain
un
substituted 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
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:
|
or
|
Description
You type
You get
Escaping pipe character as table row/column separator
{{!}}
Escaping pipe character as a plain character
|
In template calls
The same pipe protection applies as for the following example:
{{
Documentation
content
... text before ...
code
subpage-name=sandbox3
code
is equivalent to
code
sandbox link=../sandbox3
sandbox name
sandbox3
code
... text after ...
}}
We observe that
text after
is not displayed when the pipe | just before
sandbox name=
is present since
|sandbox name=
is considered erroneously to be a parameter of template
Documentation
at the same level as
|content=
is.
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
> instead of spaces.
foo
{{
#if
:||
bar
}}
foo
foobarfoo
foo{{#if:||
/>
bar
/>
}}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 foo
foo{{#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:
foo
{{
1x
|content= bar}}
foo
foobarfoo
foo
{{
1x
|1= bar}}
foo
foobarfoo
foo
{{
1x
| bar }}
foo
foo bar foo
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
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
– in particular for number formatting and padding
Manual:Template limits#Expensive parser function calls
Module:String
obsoleting
Extension:StringFunctions
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
Magic words and
Parser functions
Behaviour switches
Current
TOC
__FORCETOC__
__NOTOC__
__TOC__
Editing
__NEWSECTIONLINK__
__NOEDITSECTION__
__NONEWSECTIONLINK__
Categories
__EXPECTUNUSEDCATEGORY__
__HIDDENCAT__
__NOGALLERY__
Language conversion
__NOCONTENTCONVERT__
__NOCC__
__NOTITLECONVERT__
__NOTC__
Other
__EXPECTUNUSEDTEMPLATE__
__INDEX__
__NOINDEX__
__STATICREDIRECT__
Former
__END__
__START__
Variables
Current
Date and time
{{
CURRENTDAY
}}
{{
CURRENTDAY2
}}
{{
CURRENTDAYNAME
}}
{{
CURRENTDOW
}}
{{
CURRENTHOUR
}}
{{
CURRENTMONTH
}}/{{
CURRENTMONTH2
}}
{{
CURRENTMONTH1
}}
{{
CURRENTMONTHABBREV
}}
{{
CURRENTMONTHNAME
}}
{{
CURRENTMONTHNAMEGEN
}}
{{
CURRENTTIME
}}
{{
CURRENTTIMESTAMP
}}
{{
CURRENTWEEK
}}
{{
CURRENTYEAR
}}
{{
LOCALDAY
}}
{{
LOCALDAY2
}}
{{
LOCALDAYNAME
}}
{{
LOCALDOW
}}
{{
LOCALHOUR
}}
{{
LOCALMONTH
}}/{{
LOCALMONTH2
}}
{{
LOCALMONTH1
}}
{{
LOCALMONTHABBREV
}}
{{
LOCALMONTHNAME
}}
{{
LOCALMONTHNAMEGEN
}}
{{
LOCALTIME
}}
{{
LOCALTIMESTAMP
}}
{{
LOCALYEAR
}}
{{
LOCALWEEK
}}
Site metadata
{{
ARTICLEPATH
}}
{{
CONTENTLANGUAGE
}}/{{
CONTENTLANG
}}
{{
CURRENTVERSION
}}
{{
DIRECTIONMARK
}}/{{
DIRMARK
}}
{{
SCRIPTPATH
}}
{{
SERVER
}}
{{
SERVERNAME
}}
{{
SITENAME
}}
{{
STYLEPATH
}}
Page metadata
{{
CASCADINGSOURCES
}}
{{
PAGEID
}}
{{
PAGELANGUAGE
}}
{{
PROTECTIONEXPIRY
}}
{{
PROTECTIONLEVEL
}}
{{
TRANSLATABLEPAGE
}}
Viewed page
{{
REVISIONDAY
}}
{{
REVISIONDAY2
}}
{{
REVISIONID
}}
{{
REVISIONMONTH
}}
{{
REVISIONMONTH1
}}
{{
REVISIONSIZE
}}
{{
REVISIONTIMESTAMP
}}
{{
REVISIONUSER
}}
{{
REVISIONYEAR
}}
Affect page content
{{
DEFAULTSORT
}}
{{
DISPLAYTITLE
}}
Localisation
{{
USERLANGUAGE
}}
Statistics
{{
NUMBERINGROUP
}}/{{
NUMINGROUP
}}
{{
NUMBEROFACTIVEUSERS
}}
{{
NUMBEROFADMINS
}}
{{
NUMBEROFARTICLES
}}
{{
NUMBEROFEDITS
}}
{{
NUMBEROFFILES
}}
{{
NUMBEROFPAGES
}}
{{
NUMBEROFUSERS
}}
{{
PAGESINCATEGORY
}}/{{
PAGESINCAT
}}
{{
PAGESINNAMESPACE
}}/{{
PAGESINNS
}}
Page names
{{
ARTICLEPAGENAME
}}
{{
ARTICLEPAGENAMEE
}}
{{
BASEPAGENAME
}}
{{
BASEPAGENAMEE
}}
{{
FULLPAGENAME
}}
{{
FULLPAGENAMEE
}}
{{
PAGENAME
}}
{{
PAGENAMEE
}}
{{
ROOTPAGENAME
}}
{{
ROOTPAGENAMEE
}}
{{
SUBJECTPAGENAME
}}
{{
SUBJECTPAGENAMEE
}}
{{
SUBPAGENAME
}}
{{
SUBPAGENAMEE
}}
{{
TALKPAGENAME
}}
{{
TALKPAGENAMEE
}}
Namespaces
{{
ARTICLESPACE
}}
{{
ARTICLESPACEE
}}
{{
NAMESPACE
}}
{{
NAMESPACEE
}}
{{
NAMESPACENUMBER
}}
{{
SUBJECTSPACE
}}
{{
SUBJECTSPACEE
}}
{{
TALKSPACE
}}
{{
TALKSPACEE
}}
Escaped chars
{{!}}
{{=}}
Former
{{
NUMBEROFVIEWS
}}
Parser functions
Metadata of another page
{{
CASCADINGSOURCES:
}}
{{
#contentmodel:
}}
{{
PAGEID:
}}
{{
PAGESIZE:
}}
{{
PROTECTIONEXPIRY:
}}
{{
PROTECTIONLEVEL:
}}
{{
REVISIONDAY:
}}
{{
REVISIONDAY2:
}}
{{
REVISIONID:
}}
{{
REVISIONMONTH:
}}
{{
REVISIONMONTH1:
}}
{{
REVISIONTIMESTAMP:
}}
{{
REVISIONUSER:
}}
{{
REVISIONYEAR:
}}
URL data
{{
anchorencode
}}
{{
canonicalurl
}}
{{
canonicalurle
}}
{{
filepath
}}
{{
fullurl
}}
{{
fullurle
}}
{{
localurl
}}
{{
localurle
}}
{{
urlencode
}}
Namespaces
{{
ns
}}
{{
nse
}}
Formatting
{{
bidi
}}
{{
#dateformat:
}}/{{
#formatdate:
}}
{{
formatnum
}}
{{
lc
}}
{{
lcfirst
}}
{{
padleft
}}
{{
padright
}}
{{
uc
}}
{{
ucfirst
}}
Localisation
{{
bcp47
}}
{{
#dir
}}
{{
#FORMAL:
}}
{{
gender
}}
{{
grammar
}}
{{
int
}}
{{
#language
:}}
{{
plural
}}
Transclusion modifiers
{{:}}
{{
int:
}}
{{
msg
}}
{{
raw
}}
{{
msgnw
}}
{{
subst
}}
{{
safesubst
}}
Misc
{{
#interlanguagelink
:}}
{{
#interwikilink
:}}
{{
#isbn:
}}
{{
#special
:}}
{{
#speciale
:}}
{{
#tag:
}}
From extensions
Babel
: {{
#babel:
}}
CategoryTree
: {{
#categorytree:
}}
GeoData
: {{
#coordinates:
}}
Labeled Section Transclusion
: {{
#lst:
}} {{
#lstx:
}}
LiquidThreads
: {{
#lqtpagelimit:
}} {{
#useliquidthreads:
}}
ParserFunctions
: {{
#expr:
}} {{
#if:
}} {{
#ifeq:
}} {{
#iferror:
}} {{
#ifexist:
}} {{
#ifexpr:
}} {{
#rel2abs:
}} {{
#switch:
}} {{
#time:
}} {{
#timel:
}} {{
#timef:
}} {{
#timefl:
}} {{
#titleparts:
}}
Scribunto
: {{
#invoke:
}}
Translate
: {{
#translation:
}}
Wikidata
: {{
#property:
}} {{
#statements:
}}
See also
Help:Extension:ParserFunctions
Help:Parser functions in templates
Parser extension tags
Parser function extensions
Parser function hooks
Translate this navigation box
Retrieved from "
Categories
Extension help
Magic words
Help
Extension:ParserFunctions
Add topic