Help:Sorting

Using sortable tables
When browsing you may encounter tables that have been made sortable. A sortable table is identified by the arrows in one or more of its header cells. Clicking them will cause the table rows to sort based on the selected column, in ascending order first, and subsequently toggling between ascending and descending order.

The actual sorting process will happen on your computer using client-side JavaScript. For this reason it is only possible to use this functionality if you have JavaScript enabled in your web browser. The sorting process is also dependent on your computer and the amount of data. Sorting a very large table on a slow computer may take quite long.

Example
This is an example of a small sortable table.

Secondary key
If a column contains a value multiple times then sorting the column preserves the order of the rows within each subset that has the same value in that column (stable sorting). Thus sorting based on a primary, secondary, tertiary, etc. key can be done by sorting the least-significant key first, etc. For example, to sort the table below on the Text column, then the Numbers column, first click on the "Numbers" column heading (the secondary sort key), then the "Text" column heading (the primary sort key).

Another way to sort a table using multiple sort keys is to hold down the shift key while clicking on the column headings for the subsequent sort keys. For example, to sort the table below on the Text column, then the Numbers column, first click on the "Text" column heading (the primary sort key), then hold down the shift key and click on the "Numbers" column heading (the secondary sort key).

Creating sortable tables
Tables can be made sortable via client-side JavaScript by adding  to their top line. These tables need to be properly formatted, with the right amount of cells. Additionally you need to make sure that the headers of your column are properly indicated in the wikicode. For this the ! character is used in the table syntax.

Simple example
This is an example of the table listed in the first section and shows the typical way of how a table should be authored in order to enable table sorting.

Initial sort order of rows
When users are first presented with a table, the rows will always appear in the same order as in the wikitext. If you want a table to appear sorted by a certain column, you must sort the wikitext itself in that order. This is usually done for the first column.

Restrictions

 * Only the columns can be made sortable. Sorting of rows is not possible.

Forcing a column to have a particular data type

 * The sort mode for a column can now be manually specified by adding a   to the column header. For more info see meta:Help:Sorting.

The  attribute can be added inside the header of a column to ensure that the cells underneath are all treated as a specified type of data.

The following (case-insensitive) values are valid for :
 * for website addresses
 * for numeric internet protocol addresses
 * for dates in the ISO yyyy-mm-dd format
 * for dates in the US format (with the month before the day)
 * for numeric internet protocol addresses
 * for dates in the ISO yyyy-mm-dd format
 * for dates in the US format (with the month before the day)
 * for dates in the US format (with the month before the day)

For example:

Without  in the header, the tablesorter gets confused by the numeric titles in the first few rows into treating the entire column as numeric. This results in it wrongly sorting the non-numeric titles as zero regardless of the alphabetical ordering of their text.

Default data type of a column
If you do not specify a, the sort modes (the data types, which, in addition to the choice "ascending" or "descending", determine the sorting order) are as follows:
 * date (see also below)
 * criterion: the first non-blank element is of the form "dd-dd-dddd", "dd-dd-dd", or "dd aaa dddd"
 * order: the string abcdefghij of length 10 is positioned as ghijdeab, the string abcdefghijk of length 8 as 19ghdeab if gh>=50 (string comparison) and 20ghdeab otherwise (i.e., the assumed format is DD-MM-YYYY or DD-MM-YY), and the string "dd aaa dddd" with aaa an abbreviated month name: chronologically
 * "currency" (this mode can be useful for other data also)
 * criterion: the first non-blank element starts with $, £, €, or ¥
 * order: numeric, ignoring these symbols and all ordinary letters and commas, but not spaces; note that scientific notation cannot be used, as e and E are removed
 * numeric
 * criterion: the first non-blank element consists of just digits, points, commas, spaces, "+", "-", possibly followed by "e" or "E" and a string consisting of "+", "-", digits
 * order: after removing the commas and spaces, if any, if the string starts with a number the order is numeric according to the first number in the string (parseFloat is applied); it is regarded as zero if it is empty; in other cases (parseFloat returns NaN), the element is positioned like -∞.
 * proposed internationalisation: in German etc., treat comma as a decimal point


 * string
 * criterion: all other cases; to avoid one of the other modes, start e.g. with a hidden "&"; this can be done conveniently with, which also allows more hidden text, as a sort key; while the similar templates above are called at the end of a table element, call this one at the start
 * order: after conversion of capitals to lowercase the order is ASCII - partial list showing the order: !"#$%&'*+,-./09:;<=>?@[\]^_'az{|}~é&mdash; (see also below; a blank space comes before every other character; an nbsp code counts as a space; two adjacent ordinary blank spaces count as one; for multiple blank spaces one can use nbsps or alternate nbsps and ordinary blank spaces)

If more than one possible type matches, the first type in the above order is chosen. For example, "24-12-2007" matches as a date, so is not treated as a number. Formatting and markup tags are ignored when determining the matching type.

The sort mode is determined by the table element that is currently in the first non-blank row below the header. In the case of different data types within one column (according to the criteria mentioned above) the sort mode may change after sorting, which can give a cycle of four or even more instead of two. This is confusing and gives undesired sorting orders, so that can better be avoided. However, it can be complicated to assess whether an element can ever be at the top after any sorting operations on the same and other columns, and this can also change after deleting a row, or adding a column. Therefore it is wise to make sure that every element matches the criterion for the required data type. Using a row template this can be done very conveniently.

An alternative method of making sure the sort mode of each column is as desired, is creating a first row that determines the sort modes, but in such a way that this row is not displayed and is excluded from sorting, see below.

Examples
The script sees what the cells contains at the first data row. The sorting mode becomes numeric if the cell of the first data row contains number only (comma and full-stop used in number formatting are accepted as number). The sorting order will work properly even though other cells contains text after numbers (e.g. "200 approx"). Empty cell is treated as "zero" when sorting numerically.

It is sorted ascendingly first, and then descendingly. However bear in mind that the cell of the first data row will change accordingly after sorting. If that cell no longer contains number only after sorting, the sorting mode will change. For example, if the cell of the first data row becomes "200 approx" after sorting, this will make the sorting mode alphabetic.

The example with "a" gives alphabetic sorting; that with "e" ditto, the data are not mistaken for numbers in scientific format.

The first example demonstrates that text is positioned at zero, and that e.g. e3 for 1000 is not allowed; use 1e3 instead. It also shows that "-" should be used, not "−".

The second example shows that expressions are not sorted according to their evaluated value, but according to the first number.

The third example shows that a percentage is accepted for numeric sorting mode, but ignored in the actual sorting, so if a column contains percentages, all numbers have to be written as a percentage.

The fourth example shows again that "ca. 12" sorts at 0, as opposed to 12 with some text after it, which sorts at 12. In case such an element arrives at the top of a column, it causes alphabetic sorting mode.

Numerical sorting problems

 * Most of these problems can now be fixed by manually specifying the sort mode of a column by putting  in the column header. See the example tables below, and see also meta:Help:Sorting and the section about forcing the sort mode of a column.

Normally, text breaks numerical sorting whether the text is before or after the number. The N/A template in any cell in a numerical column breaks numerical sorting of that column.

Even when using  in the column header, text in front of a number in any cell breaks numerical sorting. Text after a number is not a problem when manually forcing the sort order of a column.

A dash, of any kind, in a blank cell breaks numerical sorting of a column. A dash is only allowed in front of a number.

A dash after a number breaks numerical sorting of a column. Therefore a range (30-40) breaks numerical sorting of a column. Click the sorting button in the first table to the right, and see that sorting does not work. One solution is to use 2 columns for a range. See and experiment with preview views of the table in 15 October 2011 global protests. Using two columns there allows one to sort by either the minimum estimated crowd sizes, or by the maximum estimated crowd sizes.

Another option is to use one column and add  in the column header. Click the sorting button in the table to the far right. For more info see meta:Help:Sorting and the section about forcing the sort mode for a column.

A plus sign (+) in front of a number does not break numerical sorting of a column. A plus sign in an otherwise empty cell can break numerical sorting of a column if the first non-empty cell going down a column consists of just a plus sign. A plus sign after a number can break numerical sorting if the first non-empty cell going down a column consists of a number followed by a plus sign. See and experiment with preview views of the table in 15 October 2011 global protests.

References after numbers currently break numerical sorting of a column. For example; compare these 2 versions of an article:. One has a reference in the "Wins" column and one does not. If  is used in the column header then references are OK after numbers.

"c." (circa, indicates "approximately") is often found in columns of numbers. It needs to be moved to a different column. Or if  is used in the column header then "c." can be put after the number.

Leading zeroes are not necessary for numerical sorting of a column. If it seems that way, then that means the column is being sorted alphabetically. Check for cells with anything other than numbers, and correct those cells according to the above rules.

Year sorting problems
Year sorting of a column works as long as no text is the first thing in a cell in the column. The year must be first. Text is OK after a year in a cell. References after the year are OK. See List of countries by income equality.

A dash, of any kind, in a blank cell breaks year sorting of a column. Dashes after the year are OK.

Unlike for numerical sorting the N/A template in any cell in a year column does not break year sorting of that column.

If there are problems with year sorting check for any cells in the column with text or a dash (of any kind) as the first thing in a cell. Remove that text or dash to return the column to correct year sorting.

Year and month
It is not possible for the default sorter to correctly sort a year and month (in text) column. The default table sorter (no manual specifying of the sort method) will not recognize the text, or text abbreviations, for months. Manually specifying the sort order by date does not help when the year is followed by the month. Try tables below to see:

Month and year
The only thing that works is adding  to the column header for columns that contain cells with both the year and the month. And the month must be first. Month abbreviations are recognized. See table below:

Any cell that only has a year (no month) breaks the date sorting of the column:

Header styling, links, and markup
Links and wikilinks are allowed in sortable headers (tooltips over links will indicate the sorting operation). Styling markup in a header may cause that column to be unsortable - see bug 31755. There are two workarounds. Styling workaround 1: use  instead of   e.g.:

Produces this sortable table:

Styling workaround 2: Add a row of empty headers above named headers (this produces non-standard appearing tables, and may not work in subsequent releases of MediaWiki). For example:

This sortable table is produced:

This allows for more columns to fit within the width of the screen. It also allows more space for the text within the header cells, since no space is taken up by the sorting icon. This is useful for tables with many columns. For example; see Look at the top table with the sorting indicator above the header, and to the side of the header. Imitate different screen sizes and resolutions by making your browser window more narrow (the middle button in the set of 3 buttons at the top of your browser window). Drag the right side of the window to get various window widths. Scroll down the page. Note that when the sorting icon is to the right of the text that the sorting icon does not wrap. So the table then requires a horizontal scrolling bar sooner than a table with the sorting icons above the headers.
 * Template:Corruption Perceptions/Corruption perceptions index.

Excluding rows from sorting
Sometimes it is helpful to exclude the last row of a table from the sorting process.

This can be achieved using  on the desired table row (line starting with  ).

Wiki markup

What it looks like in your browser

It is possible to keep multiple lines fixed at the bottom:

The first unsortable row can also be a hidden row, with each element marked with, to ensure that each column has the desired sort mode. However, this technique creates tables that have accessibility issues, as it causes problems with screen readers and text browsers. Alternatively, the entire row can be marked as hidden with, which is just as bad regarding accessibility.

Forcing proper sort type and positioning rows with a hidden sort key
One way to ensure each row is sorted appropriately is to add identical hidden rows to the top and bottom of the table. If these contain very high and very low values of the appropriate type for sorting each column, then no matter what sorting is done, one of the rows always remains at the top and one at the bottom of the table, forcing the appropriate mode for the next sort. If it is acceptable to keep identical rows fixed at the top and the bottom of a table, these can be implemented using less extreme high and low hidden sort key values. The hidden rows to force proper sort type may be unnecessary if the rows for display at the top and bottom of the table contain the right sort of values. For example, the hidden top and bottom rows in the example below can be deleted without harm, but if "(meters)" were added in the third column they would be needed.

Since the numeric sorting recognizes scientific notation, the number 9e99 and its negative are good candidates for forced numeric sorting when dealing with tables that contain large numbers. On the other hand, characters are ranked "ASCIIbetically," which places the exclamation mark (!) as the first sortable character and the tilde key (~) as the last; as such those two characters are good candidates for alphabetical sorting.

Making a column unsortable
If you want a specific column not to be sortable, specify  in the attributes of its header cell.

Wiki markup

What it looks like in your browser

Sorting with a hidden key
You can apply sorting using a key which, due to CSS, is not displayed:

Example:
 * p q r displays p q r

Javascript sorting sorts the text inside and outside the tags, without the tags themselves. A hidden key can be put at the start. Both in the case of alphabetic and that of numeric sorting the first parts determine the order. Both parts together are used to determine the sort mode, so for numeric sorting the whole should be a valid number.

Alphabetic sorting with hidden key
The key comes at the start and is separated from the displayed text in such a way that the latter does not affect the sorting order. For example, if there are no blank spaces in any key, then a blank space can be used for separation. If a single blank space is possible in a key, two nbsps can be used. For table elements for which the text to be displayed is equal to the key, no duplication is needed, of course.

If the text inside and outside the tags together is of a form that would cause a sorting mode other than alphabetic (if and when the element is at the top), a character can be appended at the end of the key to avoid this, again making sure it does not affect the sorting order by putting a space or two nbsps. This can be dispensed with if the element can never be at the top, but this can be complicated to assess as that can be caused by sorting other columns, with varying sorting modes, and it can change when deleting a row, adding a column, etc.

Instead of "display=none" another way is using a font color equal to the background, e.g. 999 gives "999 ". With this method the hidden code can be seen in selected text (e.g. with the mouse). Also the hidden text is included when copying the rendered text. The first may be an advantage or a disadvantage, the second seems only a disadvantage. A complication is also that if a user uses a background color different from the default, the specified text color may not match it; to make sure they are the same the background color can be specified also.

Numeric sorting with hidden key
If one needs to use alphabetic sort mode for numbers, one can construct a hidden alphabetic key for this purpose. The simple way of achieving this is to use template nts or ntsh.

It can also be done manually for all numbers between −1e100 and 1e100 in arbitrary precision as follows: In the following the left column shows the code for alphabetic sorting, where cryptic followed by the regular notation. The second column contains the same (hence sorting the same), but with code hidden with CSS. The third column does not contain hidden parts and uses numeric sort mode. When sorting the first or second column "more than 1e9" is positioned suitably, while when sorting the third column it is positioned like 0. Moreover, if this cell would be at the top alphabetic sort mode would be used.
 * where scientific notation is used, it is normalized such that the absolute value of the mantissa is between 1 and 10; the exponent is put first
 * scientific notation is used for all negative numbers, and all positive numbers outside some interval (below: 1e-9 to 1e9), and not inside that interval
 * where the absolute value of the exponent and/or the mantissa is a decreasing function of the number, the notation uses its complement with respect to 99 for exponents and 10 for mantissas; the code "c" is added in these cases
 * numbers 0 ≤ x < 1000 get a "+" in front
 * positive numbers in scientific notation with a negative exponent get "+0" in front
 * spaces inside the code and &-signs in front are added where needed:
 * for numbers not in scientific notation the positions of all explicit and implicit decimal points are aligned
 * for the starting position, i.e. the position of the first "-", "+", or "e", of other numbers, see the example table
 * no code should satisfy the criterion for numeric sorting mode (below we have always either an ampersand or two letters e): although this matters only for the element at the top, any element might arrive at the top due to sorting another column

Padding
Sometimes entries are padded on the left for alignment purposes. This can adversely affect how they are sorted.

Non-breaking spaces
The effect of left-padding with " " codes, which render as blank spaces, depends on the browser: in IE they are (unlike actual blank spaces) counted for sorting as leading blank spaces, so in a list of numbers with text (for which the alphabetic sorting mode applies) they could be used to equalize the number of characters before the explicit or implicit decimal separator. However, in Firefox they are ignored for the purpose of sorting.

See also w:Talk:List_of_U.S._states_by_population.

Padding with zeros
Example:
 * 000000

Formatnum can be combined with padleft:

Integer:

299,792,458 gives:


 * 299,792,458

Real:

0.000000 gives:


 * 0.000000

Dates
The simplest way to format sortable dates in a table is to use the dts template. In accordance with the manual of style, the template would be invoked using the following format:



Pleasse see the documentation of the template dts for full details on how to use this.

Issues
Currently, date sorting only works by formatting dates so they can be sorted numerically. For example:


 * yyyy mm dd

or


 * 2001 07 21

...for 21 July 2001. The "display:none" style can be used to hide a sortable numeric date before the displayed date. dts does this automatically, and is recommended in most cases.

For years BC,  can be used for -0062-09-23 (62 BC): Simply subtract the year BC from 10,000.


 * See also: 8226

Numeric sort for BC/AD years
In certain circumstances the following sort technique may be used to provide a simple intuitive numeric sort for BC/AD years which are often surrounded by qualifying text. The sort in the BC/AD column (the fourth column in the example) is forced to be numeric (just as the other columns are forced to be alphabetic) by the inclusion of the first two hidden rows. These rows contain extreme values which will mean that these rows will ALWAYS be sorted to the top and bottom of the table, regardless of which column is sorted. As the sort mechanism determines the sort type to be used by examining the first cell in the column to be sorted the hidden rows ensure that a purely numeric value is always found in the first or last cell.

The mechanisms used here are explained in the current article at: and.

With numeric sorting guaranteed all that is needed is to precede the BC/AD text with a positive or negative year number in a hidden sort key which suitably represents the cell text. As this number will be the first thing the sort code sees it will sort it as a number, in the order, or the opposite. Once the sort type is fixed at the start of a sort the presence of alphabetic values in subsequent rows is ignored. The sort is done numerically on the first text in each row. The detail of the exclamation mark after the positive/negative year number in the hidden sort key is to clearly mark an end to the number which the sort mechanism must consider. In certain circumstance, if the exclamation mark is not present, and the hidden sort key is immediately followed by another number, that number may be treated as a continuation of the hidden sort key number, to produce an incorrect sort.

The tables in the article: List of cities by time of continuous habitation have been modified to used this sort. The following example is modified (maintaining some defects in the content which will need resolving in the original!) from one of those tables:

Keeping some rows together
Partly hidden keys can be used to keep certain rows together. The original mutual order of these rows is preserved.

Example where this is the case for the rows about the Netherlands:

Limitations
Javascript sorting may not work properly on tables with cells extending over multiple rows and/or columns (however, sorting of columns up to and including the first with colspan does not seem to be affected). Also, while cells can be empty, they should not be missing at the end of a row. In these cases sometimes the table gets messed up when attempting to sort, while other times some of the sorting buttons work while others don't.

Colspan workaround
To allow sorting, the formal number of cells in each row should be equal (if not all columns are made sortable this should apply at least for the number of cells up to and including the last sortable column). However, with a CSS hack the number of cells shown in a row can differ from the formal number of cells. For example, two formal cells can be shown as one by specifying a width for the first column, shifting the contents of the second cell to the left, increasing its width by the same amount, and hiding the cell border that would normally be visible. Hidden keys can be used to control, for sorting with respect to each column, how this row should be sorted.

Example:

This can be combined with the method of "keeping some rows together" demonstrated above. For an example of an application of this, consider a table of three columns where the third column would make the table too wide, such as a column of miscellaneous details. These details can be put in separate rows, each staying below the corresponding row when the table is sorted.

Example:

A table row template makes this technique less cumbersome to apply, see e.g. List of furry conventions, w:Template:Furry-con-list-start and w:Template:Furry-con-list-entry.

Controlling sorting and display
Text undesired for sorting but needed for display:
 * In numeric sorting mode, text breaks numerical sorting whether the text is before or after the number. Sorting then becomes alphanumeric. Empty cell is treated as "zero" when sorting numerically.
 * In date sorting mode, this text needs to be put in a separate column; in the case of a cell containing a range of dates or numbers (e.g. from .. to ..), text in surplus of what is required for sorting is put in the extra column. If the first part of the text is used for sorting, then the extra column needs to be the following one; conversely, if the last part of the text is used for sorting, then the extra column needs to be the previous one; depending on the table format, this dividing of an item over two cells may look ugly.
 * In alphabetic sorting, any footnotes etc. do not require a separate column; they can simply be put at the end of the element.

Text undesired for display but needed for sorting:
 * can be put as hidden text in the column to be sorted

Combining the two, we can have displayed text independent of text used for sorting, by fully hiding the latter, and fully putting the former in a separate column (in date sorting mode and numeric sorting mode) or in the same column after the hidden text (in alphabetic sorting). Fully putting the displayed text in a separate column may look ugly if it is not done consistently for a whole column, but only for elements that require this (e.g. if most entries in a column are single numbers, but some are ranges).

Static column
A static column, e.g. with row numbers, can be obtained with two side-by-side tables with for each row the same height set in both tables:

The style can be adjusted to make it appear as a single table. If for some row the height of that row is too small for the text in a cell on one of the sides, the browser increases it, and there is no longer a match.

A static column can also be created by enclosing the sortable table with the templates static column begin and end. See the documentation of static column begin for details.

Maintaining tables sorted by rank
Tables in rank order may become difficult to maintain as the list is updated. People keep adding more rows to the table, but fail to renumber all the ranks each time a new one is added, because it seems too laborious.

One way to resort a table that has been updated is to copy the table (in wiki markup), remove the table code ("|-", etc.), replace the cell separation code ("||") with Tab characters using Notepad (built-in with Windows) or TextEdit (built-in with OS X) and paste the table data into Excel or Google Docs (free of charge). Type "1" and "2" in the first two rows of the rank column, select both cells and then fill out the rest by dragging down the little blue square on the lower-right corner of the second cell. Copy the table from the spreadsheet and paste it into a text editor. Finally, add back the wiki table code with the text editor.

Initial alphabetical sort versus initial sort by rank order

 * See also: Category:Tables that may need a separate rank column.
 * Note: A template that can be added to article or talk pages: Rank order. Or copy this link wikitext to talk pages:
 *  Help:Sorting 

It is a good idea to keep lists and tables in some kind of initial non-random sort order. Numbered rank order can be difficult to maintain as the list is updated, and as the rankings change. For example; in a nation list, updating the info for a single nation may require changing the rank numbers of many nations.

This discourages people from updating the list. If they do update the list, they may not bother to update the rank order. So the list becomes more and more incorrect over time. See lists of country data. See Category:Lists of countries by per capita values and Category:Lists of countries.

It can be even easier maintenance and updating in some cases if the list is initially in alphabetical order, especially if the sources and references are also in alphabetical order. Readers can be instructed to use the sort button to order a number column.

Removing a rank column (1,2,3) from a table
You can remove the rank column cells one-by-one manually, or with a quicker method. Unfortunately, none of the quick methods listed below can be done with the Notepad text editor that comes with Windows. The methods below are specific to NoteTab Light (freeware version of NoteTab). It installs quickly and easily.

If the rank column (1,2,3) is the first column, then there is a fast way to remove it. First paste the wikitext for the table list into a new page in Notetab Light. Click your mouse cursor at the top left. Do not select anything. Click on the "modify" menu, then "block", and then "delete". Enter the block dimensions in the dialog box. For long lists you will need to break up the list first into a 1-digit (0-9) rank section, a 2-digit (10-99) rank section, and a 3-digit (100-999) rank section.

To get the number of lines go to the tools menu, then "text statistics", then "more", and then scroll down to the bottom for the total number of paragraphs. Enter that number in the spot for "block rows" in the dialog box. Then click OK. The block with the rank column will be removed.

Then put back   |-    (wikitext for row) between each line. Do that via find-and-replace by replacing ^p with ^p|-^p


 * ^p is the underlying text editor code for line breaks in Notetab.
 * |-   is the wikitext for a table row.

If there are blank lines between the entries replace ^p^p with ^p|-^p

Putting a table in alphabetical order
If you also want to alphabetize the table, then first make sure the alphabetical column is the first column. To quickly alphabetize the list paste the wikitext into a new page in Notetab Light.

Click on the "modify" menu, then "lines", then "sort", and then "ascending". That will put "A" at the top and "Z" at the bottom. Put back   |-    (wikitext for row) between each line as described above. Copy the wikitext and paste it back into the article. Save the page.

Adding a separate rank column (1,2,3) next to a table

 * Note: Copy separate rank-order table (1,2,3) from here. It is 1 to 250. Use as much as is needed. See also this list of just the numbers.

There is a Bugzilla thread asking for a table option for a static rank-order column, or row numbering, as an integral part of a table. See 40634. For info on a separate rank column see below.

The rank column (1,2,3) can be removed, and put to the side of the table. See examples farther down. The rank column then never has to be updated, except to add more numbers at the end if necessary. This allows any initial sort order for the table. It allows any column to be sorted later via the sort button.

Click the sort button above a column in the main table on the right. Note that the rank column does not change. This allows one to see rank order for any data column. A note can be added above the table; something like this:


 * See the rank order of any data column by clicking the sort button [[File:Sort both.gif]] above any data column.

A separate rank column only works when there is only one line per row in the main table. So tables with a notes column are problematic. There could be 2 lines in some rows, depending on screen width, and the length of notes. For example: List of countries and dependencies by area.

Tables that extend all the way across the page may be problematic for the same reason. In some cases this can be fixed by narrowing the headers by using breaks in the header titles:   . The number of lines in the header in the rank column can be made to match that for the main table. See examples farther down.

Tables with references in cells other than header cells are problematic. Reference numbers are in superscript, and this increases the height of the cell. So a separate rank column may not work in this case, especially when multiple cells in a column have references. See this example. Note that the rows in the rank-order column gradually stop lining up with the rows to the right as one goes down the table.

Here is how to combine a main table and a separate table with just a rank column.

The two tables are combined by wrapping them within a 3rd table that does not have a border. Look at the wikitext to understand. See examples below, and here: List of countries by public debt.

{|
 * - style="vertical-align:text-top;"




 * }

The number of lines in the rank header cell should match the maximum number of lines in header cells in the main table.

2 lines in header cells.

3 lines in header cells.

References in the headers in the main table add a little height to the header cells. This additional height can be ignored since it does not effect alignment much. Or a reference note can be added to the rank column to make it line up more perfectly with the main table.

{|
 * - style="vertical-align:text-top;"


 * }
 * }

Sorting the wikitext of a table
One difficult way to sort the wikitext itself by a certain column is to use the following 'trick'. You can make an auxiliary sortable table containing the wikitext for the original table, and sort it. You can then replace the original wikitext by this sorted wikitext.

Unfortunately this method is anything but quick and easy. The heavy modifications needed make it unsuitable for larger tables, while small tables generally are sorted faster by hand.

Example:

Original table:

Auxiliary table:

Now you can sort the above table, and copy the rendered text to the edit box. After deleting the "header" line, this renders as a new defaultly sorted table:

Alphabetic sorting order
Sort the following table to see an example of the alphabetic sort order. Note that sorting is case-insensitive: the two-character entries such as A1 demonstrate that A and a are at the same position.

Browser issues
With older versions of Safari a table can only be sorted by the first column: all sort buttons have the effect that only the leftmost one is supposed to have.

JavaScript source
The JavaScript code is found at /skins-1.5/common/wikibits.js. In addition, MediaWiki:Common.js adds and overrides some code.