Controlling output format

You can select one of several default formats or define your own format ; output can be grouped in columns or rows.

General approach to output formatting
The general approach to output formatting is two-fold:


 * 1) There are a couple of simple predefined output formats which generate lists of articles
 * You will understand their meaning directly from reading
 * 1) There is a mode called "userformat" which puts complete control into your hands
 * This is somewhat complicated.

While the standard output formats are meant to be used for fast generation of simple page lists, the userformat approach aims at transcluding contents from other pages and requires some effort to understand. There is a system of three tags which are used to enclose (a) the whole output, (b) each item, (c) each transcluded section of an item. A fourth tag is used to separate values between items of one section which occur more than once.

We assume that we have two documents which use templates x and y with varying arguments; while x is being used once within each document, y is used several times. In very short notation the structure might look as follows: A: x(a) y(3) y(5) B: x(b) y(4) y(1) y(2)

The following DPL parameters are used to define a set of tags which are used to construct the output: The arguments of the above statements can contain references to %VARIABLES%. So sec-1-start might contain a reference like  to output the page name. See format for more details on variable substitution.
 * = liststart,itemstart,itemend,listend
 * = sec-1-start,sec-1-end,sec-2-start,sec-2-end, .., ..
 * = multi-sep

Now think of the following page inclusion statement: includepage={x}.dpl,{y}.dpl

The output will then look like this: liststart itemstart sec-1-start x.dpl(a) sec-1-end sec-2-start y.dpl(3) multi-sep y.dpl(5) sec-2-end itemend itemstart sec-1-start x.dpl(b) sec-1-end sec-2-start y.dpl(4) multi-sep y.dpl(1) multi-sep y.dpl(2) sec-2-end itemend listend

Assuming that the tags (, , etc.) contain wiki syntax for table definitions and multi-sep defines a horizontal line, the output might look like this:

+--+-+ |      |          | y.dpl(3) | | A   | x.dpl(a) |      | |     |          | y.dpl(5) | +--+--+--+ |      |          | y.dpl(4) | |     |          |      |  |  B   | x.dpl(b) | y.dpl(1) | |     |          |      |  |      |          | y.dpl(2) | +--+--+--+

In some situations, however, you may want to create an output table where each of the calls of template y is used to create a separate output row. Using a sortable table you could then easily rearrange the output.

+--+-+      +--+-+  |  A   | x.dpl(a) | y.dpl(1) |       |  B   | x.dpl(b) | y.dpl(1) | +--+-+      +--+-+  |  A   | x.dpl(a) | y.dpl(2) |       |  B   | x.dpl(b) | y.dpl(2) | +--+-+      +--+-+  |  B   | x.dpl(b) | y.dpl(3) |       |  A   | x.dpl(a) | y.dpl(3) | +--+-+      +--+-+  |  B   | x.dpl(b) | y.dpl(4) |       |  A   | x.dpl(a) | y.dpl(4) | +--+-+      +--+-+  |  B   | x.dpl(b) | y.dpl(5  |       |  B   | x.dpl(b) | y.dpl(5) |  +--+-+       +--+-+ There is a special parameter called   which you can use to mark one section of your   statement as "dominant" (in our example:   as {y}.dpl is the second argument of our   statement). You can only have one dominant section in a DPL statement. Marking a section as "dominant" only makes sense if you have multiple calls of the same template (or multiple chapters with the same heading) in your documents. Each piece of content in the dominant section will generate an individual output row with the values of all other columns being repeated.

As all of the above is not very easy to understand there are additional DPL commands which make it fairly easy to create tabular output.

mode
Syntax:

modename can be one of:


 * unordered: outputs an unordered list &mdash; HTML tag "ul" &mdash; (default)
 * ordered: outputs an ordered list &mdash; HTML tag "ol"
 * subpage: outputs a nested unordered list with each subheading containing pages that exist under it. &mdash; HTML tag "ul"
 * none: outputs a list using newlines and HTML tags "br" to separate each item
 * inline: outputs a list using symbol defined by the  parameter to separate items.
 * category: outputs resulting articles in a way category-pages are shown (you have to use  with this option!)
 * gallery: outputs resulting content in a gallery, as would be done with the HTML tag "gallery"; it is likely this would be used with files, which must be explicitly requested using the namespace option


 * userformat: will leave output control completely to the user; see parameter  and  ; in this mode DPL2 will offer built-in variables which must be referenced in the output format description provided by the user. mode=userformat is quite important to have complete control over the output.

For advanced use of DPL it is important to understand. Note that this mode is automatically implied when  or   are used.

Example2
This list will output pages that have 'frica' in their name; pages will be ordered by their name regardless of category, the output will be shown in category style (i.e. with chapter capitals)

Related DPL extension variables:.

inlinetext
Syntax:

, with wikitext some wiki text default is '&amp;nbsp;-&amp;nbsp;' except for mode=userformat where inlinetext is empty by default.

If you want normal "breaking spaces" (and not the NON-breaking spaces) you should use '&amp;#32;-&amp;#32;'.

Extra whitespaces are stripped by DPL from the beginning and end of wikitext. If you want to show one or multiple spaces, use one or multiple ' ', or use 'nowiki' tags like in '&lt;nowiki&gt; - &lt;/nowiki&gt;' which has same effect as '&amp;nbsp;-&amp;nbsp;'.

Example:

This list will output pages that have shown like Item1 • Item2 • Item3 • ...

listseparators
Note that format= automatically implies mode=userformat whereas  does not.

format
NOTE:  is an alias for   which does not automatically imply.

NOTE: The  command is very flexible, but can be complicated to use. Creating tabular output is better done with the table command.

Syntax:

'Startall', 'Start', 'End' and 'Endall' are wiki tags used to separate the list items.
 * 'Startall' and 'Endall' define an outer frame for the whole list.
 * 'Start' and 'End' build an inner frame for each article item.

Because wiki syntax depends on newline characters the string  must be used to explicitly insert newline characters into the output.

As we want to be able to control output completely we reference article names and other possible output by special symbols:


 * = the current article sequence number (starting from 1)
 * = the name of the page (including namespace)
 * = the internal unique numeric ID of the arcticle page
 * = the physical path to an image (based on hash values, e.g. 5/5d/myImage.jpg)
 * = the name of a page which was used within the selection criteria (only applies to  and  )
 * = the name of an image which was used within the selection criteria (only applies to )


 * = the title of the page (without the namespace)
 * = the namespace of the page


 * = the usage counter (requires )
 * = the article size (requires )
 * = a font size number which is based on the article size (logarithm of square root of counter)
 * = a font size number which is based on the usage counter (currently this is the logarithm of the usage counter)
 * = similar to, but based on the logarithm of the square root of the usage counter.
 * = the change log message of a revision; only accessible if the DPL query is based on revisions.


 * = the date selected, eg. ; requires   or similar; the formatting of the date can be influenced using
 * = the user who changed the document last; requires


 * = the user who made a contribution; requires
 * = the number of bytes changed; requires
 * = an asterisk bar to indicate the amount of change; requires


 * = a pipe-separated list of links to all categories to which the article belongs (requires )
 * = a bullet point list of links to all categories to which the article belongs (requires )
 * = a comma-separated list of all categories to which the article belongs (requires )


 * = the name of the revision of the article; this will only be available if you made a selection based on revisions


 * = the external hyperlink found as a consequence of the linkstoexternal statement

These symbols will be replaced by the corresponding values if they occur within 'Start' or 'End' or within the corresponding tags of the  parameter.

In addition there are some symbolic variables which can ONLY be used in  and  :
 * = number of articles in the result set
 * = total number of articles in the result set, regardless of count limits; will only be calculated if used.
 * = the current DPL version


 * = contains the amount of time (in seconds + milliseconds) spent within DPL; this can be helpful if you observe slow response times for wiki pages that contain DPL statements. Example: 2 (2009/06/13 09:27:43) would mean that DPL spent two seconds of the whole response time, starting at the time given in brackets.


 * : These variables contain namespace and title of the first / last article in the result set; the information is intended to be used for page scrolling
 * : The value of this variable is set by the URL parameter . It is passed to the scroll helper template which uses it to produce its links for scrolling.

For example the classical default output of DPL2 can also be produced with the following statements:

Example:

Note that a bullet point list in wiki syntax is defined by a "*" at the beginning of a line -- therefore we have to use a special symbol '\n' or '¶' to refer to the beginning of a new line of wiki text. Replace the "*" by a "#" and you will get a numbered list. 'Startall' and 'Endall' are empty (note that we start with a comma, note the two commas at the end), the 'Start' tag is used to create a new line with an initial "* " followed by the page name, written as a link. That's all.

Creating a top-five hitlist with access rates and bold article names of varying size could be done like this:

You can also use HTML syntax for the tags, although this is discouraged.

Now let us create a table using wiki syntax:

We use 'Startall' to define the table header and 'Endall' for the footer. Each article is presented in a table row using wiki syntax for table layout.

Syntax
or

''Please note that the semantics of this parameter have changed with version 0.9.6!. When upgrading to 0.9.6 it will probably be necessary to change the  statements.

In the first syntax variant, specify pairs of tags which correspond to the  statement. and  are HTML strings or wiki tags which will be put around each transcluded section (see  ).

In the second syntax variant, specify just one element which will then be used as 'StartN' for all sections; in this case the second tag will be empty for all transcluded sections.

Symbolic replacements of  etc. take place as described in. In addition, the symbol  can be used to refer to the section found (works only for chapter headings).

If the same section occurs more than once in an article (or an article includes the same template more than once) all such occurences will be transcluded as a block and the  tags will only be put once around the whole block (but see  ).

Example
Which produces,

Use  to define a table with three columns and put a link to the article in the first column of each row. Use  to add more columns for each section found. There are two pairs for each transcluded section; the first element of each pair is a linefeed and a pipe (which define a new column in the table) and the second element of each pair is empty. Have a careful look at the '¶' symbols ('\n' can be used as an alternative). They always appear before a wiki syntax element which must be placed at the beginning of a new line. Thus, make sure that the wiki parser will understand them. Note: if an article does not contain a section named "fruit", it will result in an empty cell in the table.

As mentioned above, a single element can be used in the  statement in order to apply this as a start tag to all transcluded sections; so it could have also been written:

Example 2
Assuming that the chapters on fruit and color contain long texts, they can be truncated to 50 or 100 characters. A link which refers directly to those chapters will be generated automatically if needed.

multisecseparators
Syntax:

The tags correspond to the transcluded section (see ).

Symbolic replacements of  etc. take place as described in. In addition the symbol  can be used to refer to the section found (works only for chapter headings). It will give you the precise name of each heading even if you used a regular expression (double ##) in the include statement.

If an article uses the same template more than once you will get all references with "sepN" as a separator.

Example:

See also Test article structure.

dominantsection
Syntax:

between 1 and the number of arguments in your  statement

If there is only 0 or 1 piece of contents for the dominant section you will see no difference from normal DPL behaviour.

Example:

See the explanations at the top of this document to understand the meaning of dominantsection.

Note: Using  together with   may lead to strange result formatting.

table
Syntax:

tableatr, linkheader, (column headlines) ..

The  statement is a shortcut which implicitly sets certain values for other DPL parameters, namely ,   /  ,  , and.

The layout is less flexible than the individual use of all of the above parameters but will probably be sufficient in many cases, especially when used together with tablerow.

If you use  in a DPL statement, it does not make sense to use one of the other options mentioned because their values will be overwritten without notice. There is one exception of this rule: It can make sense to specify the THIRD argument for  in combination with. Therefore this parameter is NOT overwritten by the  command. The third argument can be used to output meta data like %COUNT%, %USER% etc. as columns in an output table. If you want to do so, the third parameter must contain wiki syntax for output columns like this: include = {some template}:parm1,#some heading table  =,,tplparm,chapter,#hits format =,,\n%COUNT% Do not forget to escape the '|' symbol if your DPL statement uses parser function syntax. You will get a table which contains template parameters, chapter contents and the usage counter as a third column. Meta data can only be placed AFTER normal contents as we use the THIRD parameter of the  statement.

The use of  requires an   statement which should, for reasons of readability, directly precede the   statement). Each argument of the   statement will produce one or more columns in the output table described in the   statement.

 expects a comma-separated list of parameters:


 * The first parameter will be used to describe general parameters for the table
 * it is recommended to make a CSS reference here, using something like  or   if mytable is defined in the Mediawiki:Common.css document.
 * is the default value. Use double-quotes to specify multiple classes, e.g.,.


 * The second parameter is the headline for the first column.
 * The first column will automatically contain a reference to the article, so something like Article should be o.k.
 * is the default value.
 * if you use a single  (dash), the column with the hyperlink to the article will be suppressed. You can supply a hyperlink to the article in any other column if you use   within a phantom template.


 * All subsequent parameters are column headings which correspond to the arguments of the  parameter. Note that if you call a phantom template (like  ) in the include statement, you will have to provide as many headlines as the phantom template produces columns.


 * will be set to userformat


 * will be configured to produce wiki syntax which defines a table


 * will be configured to produce wiki syntax which creates a table row. The first column will always contain a hyperlink to the article of the query result (except you set the link header to '-' as described above.


 * will be configured to produce wiki syntax which creates another table row for multiple occurencies of the first  argument. For all other arguments a linebreak will be used if we are dealing with template parameters and a horizontal separation line will be used when dealing with chapter contents. The background for this is the following: If you have an article which calls the same template several times, you may want to have a table where each template invocation becomes a row in your table.

When using phantom templates (i.e. templates which are called during DPL execution instead of the original template) they must be written to produce output according to wiki table syntax. When entering such a template we are already at the beginning of a column (i.e. a preceding line with a  has already been put into the output stream). So start directly with the contents of the first column. To add more columns use a  in a separate line. Example: some output for the first column: |  some output for the next column: |  some output for the next column:

It may sound complicated, but is a huge improvement compared to the native use of  and.

A typical DPL statement using the  parameter would contain: include =                         #Chapter X,{T1}:parm1,#Chapter Y,{T2}.dpl table  = class=sortable, Article,      X   ,     t-p  ,     Y    , T2-a, T2-b Note that we have written the above statement in a way to show the correspondence between  and. You can see the first two parameters which define the table characteristics and a headline for the hyperlink to the article. Then follow headlines for each argument of. Note that there are TWO headlines which correspond to the last argument of the  statement (assuming that Template:T2.dpl outputs TWO columns). Template:T2 itself might have more or less than 2 arguments -- it only matters how many columns are output by Template:T2.dpl).

Now look at the examples

tablerow
Syntax:

coldef, ..

Where coldef contains wiki code which uses the symbol '%%' to refer to the corresponding element of an  statement.

The  statement (which must be used as a prerequisite for  ) cares for the basics of table generation. So, when you define a column definition, you only need to specify the code for the field contents itself. You can start with field attributes like "bgcolor" or skip them. You can add a leading "\n" to make sure that the field contents is displayed correctly if it contains wiki syntax that depends on linebreaks (e.g. enumeration list). You must specify all columns. i.e. you must have as many entries in the  statement as there are columns in your table. Skipping a column would suppress output for that column completely. The  command is best explained by an example (you can also find this example on the title page of this wiki):


 * We select some African countries, we exclude Sudan
 * We acquire a piece of meta data about the articles we will find (in our case the usage counter)
 * We include two named parameters from a template call, Infobox Country or territory
 * We include a short text passage from the text before the first chapter
 * We use the third parameter of  to output the usage counter (we could even highlight values above 100 here, for example)
 * We highlight usage values above 300
 * We define a standard wiki table with the article name in the first column (named "Country")
 * We define column headers for transcluded contents and for the usage counter
 * We care for right alignment of the numerical values and define a bgcolor for the text

tablesortcol
Syntax:

number

number is the position of the column that shall be used as sortkey when the result is initially displayed.
 * column numbering starts with 1;
 * tablesortcol = 0 means do not sort; this is the default.
 * Negative numbers are used to sort in descending order; e.g. -3 would sort according to the third column in descending order.
 * Note that the rest of the row after the selected column will also be part of the sortkey; so the contents of successive columns may serve as a secondary sort criterion if there are identical values in the selected column.
 * Also note that the whole column contents is taken; this may include hidden contents or markup sequences if you used column formatting commands. For the same reason you cannot expect numeric contents to be sorted 'numerically' - sorting will always be alphabetical.
 * you can of course use something like 'class=sortable' together with 'tablesortcol'. The difference is that ..
 * interactive sorting only works after the article has been initially displayed
 * interactive sorting tries to guess the content type of a column and sorts according to that (date, number, string)
 * If you do not use  the output order of your table rows will depend on the sort order by which the articles were analysed. That order depends on other DPL commands like  . The default is "alphabetically by title". So, without 'tablesortcol' you get the tablerows in alphabetical sequence of the article names where they come from. With 'tablesortcol' you can order them by the column contents itself.

headingmode
Syntax:

modename can be one of:
 * none &mdash; headings are not displayed, no heading &mdash; (default)
 * unordered &mdash; outputs an unordered list &mdash; HTML tag "ul"
 * ordered &mdash; outputs an ordered list &mdash; HTML tag "ol"
 * definition &mdash; outputs a definition list &mdash; HTML tag "dl"
 * h1 &mdash; outputs sections &mdash; HTML tags "H1"
 * h2 &mdash; outputs sections &mdash; HTML tags "h2"
 * h3 &mdash; outputs sections &mdash; HTML tags "h3"
 * h4 &mdash; outputs sections &mdash; HTML tags "h4"
 * h5 &mdash; outputs sections &mdash; HTML tags "h5"
 * h6 &mdash; outputs sections &mdash; HTML tags "h6"

Example:

This list will output pages that belong to one of the categories Africa, Europe in a list similar to this (HTML source), with pages 1 and 2 in the Africa category, Page 1 in the Europe category as well (in fact, the titles are replaced with the appropriate links):  Africa   Page1 Page2   Europe   Page1</li> </ol> </dd> </dl>

Headingmode can be used with multi-column output but the length of the columns may in this case vary more than you would expect.

headingcount
Syntax:

default is

listattr
Syntax: Examples:

itemattr
Syntax: Example: see listitemattr.

hlistattr
Syntax: Example:

See also hitemattr.

hitemattr
Syntax: Example:

userdateformat
Syntax:

The formatstring may contain letters like "y,Y,m,M,d,D,h,H,i,I,s" for year, month day. Other characters are printed as they are. See the documentation for php function date for more details. The "userdateformat" applies to all date/time fields, see the parameters: addeditdate,addpagetoucheddate,addfirstcategorydate

Example:

userdateformat=Y-m-d (D)

Default:

By default DPL uses "Y-m-d H:i:s" to display date and time. Note that MediaWiki stores all dates/times in UTC format. When displaying a time stamp DPL will translate it according to
 * 1) the timezone preference (difference to UTC/GMT) given by the user in his user settings
 * 2) if no preference is given and for all anonymous users the local time on the server will be used.

So you will either see a time based on your local time (browser based) or based on the timezone in which the wiki server is running.

The same kind of translation applies to dates you specify when selecting articles by revsion date/time.

shownamespace
Syntax:

Example:

This list will output all Talk pages in, listed without the Talk: prepended to page names.

Note that in "mode=userformat" there is a different way to decide whether you want to output the title with or without namespace. In mode=userformat two built-in variables are provided which contain the page name including the namespace (%PAGE%) and the base title name (%TITLE%).

escapelinks
Syntax:

Note: You can use this parameter to show images; another way to do this is to use the gallery extension in combination with DPL; there is an example for this on the dpldemo website.

titlemaxlength
Syntax:

replaceintitle
Syntax:

The "search for" argument must be an expression which can be used in a php preg_replace function call.

Example:

to remove the string "demo" in article names, you must write
 * replaceintitle=/demo/,

Note that standard regexp rules apply. The regexp must start with a non-alphanumeric character -- but not with a backslash! It is good habit to use a '/' if this character is not needed within the regexp itself. Read the php manual to understand the details of regular expressions.

columns
Syntax:

Example:

Simply displays all articles and subcategories in Category:Test in 3 columns ( is used to make the table width 100%).

More complex example:

In  the outer tags from   will be repeated for each column.

The output will contain a list of the 12 largest articles on Africa, arranged in three columns. Each column consists of a table which has itself three columns: rank, article name and size. Unfortunately, it doesn't work...

rows
Syntax:

In "mode=userformat" the outer tags from "listseparators" will be repeated for each column. Thus you can create long lists where the table heading is repeated from time to time. Example:

The output will contain a list of the 12 largest articles on Africa, arranged in two rows (of 6 lines each). Each row consists of a table which has itself three columns: rank, article name and size.

rowsize
Syntax:

In "mode=userformat" the outer tags from "listseparators" will be repeated after each group of "rowsize" output lines. Thus you can create long lists where the table heading is repeated in regular intervals. Example:

The output will contain a list of all articles on Africa. After each group of 20 entries (article names) the table heading will be repeated. It may be useful to set the width of the column with the article names explicitly, so that the tables in each row have equal width.

rowcolformat
Syntax:

Example:

There will be more space around the columns than normal. See  above for another example.

The ideal way to use  is to assign a CSS class to your DPL table which has been defined in your mediawiki:Common.css article.

Example:

In your Common.css article you might have written something like